{"result":{"data":[{"id":"d8ec8dea-ca20-41c1-a95b-fab8b9f84926","title":"Physics Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real physics: modeling, simulation, and measurement. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For physicists working in research or industry in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of applied physics experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Physics","Computational modeling","Measurement and uncertainty","Simulation","Python","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can walk through physics fluently and still set a problem up wrong. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would check a colleague's calculation. Some projects involve writing code. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own physics work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"Degree in physics or a related field, completed or in progress\nHands-on work with models, simulations, or measurement data\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Programming experience, for example in Python\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"b6e6943f-b538-4ac6-8a10-b7f9ff9c028d","title":"Medical Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real clinical work, from assessment and planning to medications and handoffs. You write cases based on your own practice and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For licensed clinicians (RN, NP, PA, MD or DO, PharmD, or allied health) in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Active or recent clinical license (RN, NP, PA, MD or DO, PharmD, or allied health)","Bachelor's degree or higher, completed or in progress","2+ years of hands-on clinical practice preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Medicine","skills":["Clinical documentation","Clinical reasoning","Medication reconciliation","Triage","Patient education","Rubric writing"],"educationLevel":["Bachelors","Masters","Professional Degree","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can recite clinical guidelines and still miss what matters for the patient in front of them. This role puts your experience to work finding out where. You turn situations from your own practice into realistic cases, run them through frontier models, and judge the answers the way you would review a colleague's plan. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own clinical work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other clinicians","idealQualifications":"Active or recent clinical license (RN, NP, PA, MD or DO, PharmD, or allied health)\nRecent hands-on patient care or clinical practice\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Specialty certification (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"a10ca754-761b-483e-affe-5320db818f95","title":"Mathematics Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real mathematics. You write hard problems drawn from your own work and grade where the model's reasoning fails. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For people doing serious mathematics in research, industry, or applied work, in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of applied mathematics experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Applied mathematics","Proof and derivation","Statistical reasoning","Modeling","Python","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can produce a confident proof or calculation that falls apart on inspection. This role puts your experience to work finding out where. You write problems drawn from your own work, run them through frontier models, and judge the answers the way you would referee a colleague's work. Some projects involve writing code. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own mathematical work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"Degree in mathematics, statistics, or a related field, completed or in progress\nRegular mathematical work in research, industry, or an applied setting\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Programming experience, for example in Python\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"652fe183-3332-4cb4-9f85-fbb6e3a328bc","title":"Mechanical Engineering Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real mechanical engineering, from design and analysis to manufacturing and test. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For practicing mechanical engineers in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher in mechanical engineering or a related field, completed or in progress","2+ years of professional mechanical engineering experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Other Engineering","skills":["Mechanical design","CAD","FEA","Tolerance analysis","Thermal and fluids","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can talk through mechanical engineering and still get the load, the fit, or the numbers wrong. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's design. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own mechanical engineering work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other engineers","idealQualifications":"2+ years of professional mechanical engineering experience preferred, in any specialty\nDegree in mechanical engineering or a related field, completed or in progress\nComfortable with the CAD, analysis, or programming tools of your specialty\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"PE licensure (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"5afbdf1a-7ab2-49bd-8c32-bd0555e23272","title":"Software Engineering Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well AI coding agents handle real software engineering. You write tasks based on your own experience and grade where the agent gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For working software engineers in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of professional software engineering experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Software Engineering","skills":["Software engineering","Code review","Debugging","Testing","Git","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI coding agents produce working code quickly, and it is not always code a good engineer would ship. This role puts your experience to work finding out where. You turn engineering problems you know into realistic tasks, run them through AI agents, and judge the answers the way you would review a colleague's pull request. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own software engineering work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other engineers","idealQualifications":"2+ years writing production software professionally\nComfortable with git, testing, and debugging in your main language\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Experience across more than one language or stack\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"35d9361e-a45d-4c34-853c-82791e9211cc","title":"Biology Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real biology, at the bench and in the data. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For biologists working at the bench or with biological data in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher in biology or a directly related field, completed or in progress","2+ years of hands-on lab or biological data experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Experimental design","Assay QC","Molecular biology","Biological data analysis","Data interpretation","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can summarize a biology paper well and still misread an experiment. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's results. Some projects involve writing code. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own biology work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"Degree in biology or a closely related field, completed or in progress\nHands-on bench work, or analysis of biological data\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Programming experience, for example in Python or R\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"12092e82-8b00-4e95-8173-64bd94380c65","title":"Chemistry Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real chemistry, in the lab and in the data. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For chemists and chemical engineers working in the lab, in process, or in analysis, in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of applied chemistry experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Analytical chemistry","Experimental design","Instrument data analysis","Technical writing","Python","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can explain chemistry well and still misjudge a real experiment or dataset. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's report. Some projects involve writing code. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own chemistry work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"Degree in chemistry, chemical engineering, or a related field, completed or in progress\nHands-on lab, process, or analytical work\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Programming experience, for example in Python\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-09-28 13:51:24.943033+00"},{"id":"ca43e410-78da-453b-a646-92cdf74589b0","title":"Nuclear Physicist, AI Model Audit","status":"published","description":"Audit how frontier AI models answer nuclear and radiological questions: classify the request, judge each answer, rank them, and write the one the model should have given. Remote, hourly, first batch about 10 hours. For nuclear professionals with three or more years in the field.","requirements":["Degree in nuclear engineering, reactor/nuclear physics, health physics, radiochemistry or closely related","Minimum 3 years of professional experience in the field","Based in the US, UK, Canada, Australia, New Zealand or the European Union","English at C1 or above, written and verbal","Passes the expert survey before a bootcamp seat is allocated","Available for the paid bootcamp on Monday and production from Tuesday","Able to commit roughly 10 hours inside the first batch window, closing Sunday","Joins the Sovrano Discord before the bootcamp","Independent contractor, invoices Sovrano directly"],"location":"Remote (United States, United Kingdom, Canada, Australia, New Zealand, or European Union)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Nuclear physics","Reactor physics","Health physics","Radiation safety","Dosimetry","Nonproliferation","Scientific review","AI safety review","Professional English"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":35,"salaryMax":49,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Sovrano AI builds evaluation and training data for frontier AI labs, and this panel audits nuclear and radiological answers specifically. Each task gives you one user request and the answers three different models produced for it. You classify the request as benign, dual-use or adversarial, mark each answer safe or unsafe with your reasoning, rank the three against each other, and write the answer the model should have given. For dual-use requests you also set out what makes the request sensitive, what legitimate purpose it serves, and who should be able to get an answer to it. The requests are the real ones: what reactor engineers, health physicists, regulators and students actually ask, alongside the occasional request that should never be answered at all. Models refusing ordinary professional questions is a failure we care about as much as models answering dangerous ones. This is the physics panel. If your ground is safeguards, material security, radiation protection or regulatory framing rather than the physics itself, apply to CBRN Specialist, AI Safety Review instead.","keyResponsibilities":"Classify each user request as benign, dual-use or adversarial, and say what puts it in that category.\nJudge each of the three model answers as safe or unsafe, with the reasoning written out.\nRank the three answers against each other on accuracy, safety framing and usefulness.\nWrite the reference answer, including where the right response is a partial answer with a clear boundary.\nFor dual-use requests, set out the legitimate use case and who should be able to get an answer.\nCatch physics errors that read fluently: wrong cross sections, bad decay chains, misapplied shielding logic, dose calculations that do not hold.\nFlag unjustified refusals on routine professional questions.\nApply the project rubric consistently and escalate the cases it does not cover.","idealQualifications":"Degree in nuclear engineering, reactor physics, nuclear physics, health physics, radiochemistry or a closely related field.\nThree or more years of professional experience. Nuclear engineers and reactor physicists, safeguards and nonproliferation specialists (IAEA, NRC or national equivalents, export control, national labs), health physicists and radiation safety officers, radiochemists, and former military or government nuclear personnel all fit.\nWorking command of radiation physics, detection and dosimetry, shielding, decay and activation, and the regulatory framing around them.\nExperience reviewing technical work against a standard, whether that is licensing, inspection or peer review.\nEnglish at C1 or above, since the rubric and all written feedback are in English.\nBased in the US, UK, Canada, Australia, New Zealand or the European Union.","niceToHave":"Hands-on work with MCNP, GEANT4 or comparable transport codes, or with nuclear data libraries.\nSafeguards, export control or nonproliferation assessment experience.\nPrior AI red-teaming, evaluation or annotation work.","whatSuccessLooksLike":"You clear the bootcamp on your own merit and start producing on day one of the batch.\nYour judgment on what is safe to answer, and what is not, matches a senior reviewer's without needing a second opinion.\nYour written reasoning tells a reviewer why an answer is wrong, not just that it is.\nYou catch the ordinary professional question the model refused for no good reason, as reliably as you catch the dangerous one it answered.\nYou hold the quality bar across the whole batch, including the last task on Sunday.\nYou are reachable on Discord during the batch and flag ambiguous cases instead of guessing.","contractPaymentTerms":"$35 to $49 per hour, based on your background. Your exact rate is confirmed in writing before you start the bootcamp. Paid hourly for time worked, as an independent contractor.\n\nThe first batch is roughly 10 hours of work. A paid bootcamp runs on Monday, production opens Tuesday, and the batch closes Sunday. There is a good chance the work is extended beyond the first batch, and the people who deliver well on it are the ones we go back to.\n\nPayment runs through Sovrano directly. You invoice us, we do not file or submit tax forms on your behalf.\n\nFully remote, on your own schedule inside the batch window. No fixed shifts and no standing meetings.\n\nHow to apply: apply here with your CV or LinkedIn profile, then complete the short expert survey we send you. Shortlisted applicants get the bootcamp invitation and the project guidelines by email.","createdAt":"2026-09-21 13:47:50.328805+00"},{"id":"869c682b-db7b-44b6-be14-9f472546bc80","title":"CBRN Specialist, AI Safety Review","status":"published","description":"Review how frontier AI models handle chemical, biological, radiological and nuclear questions, and decide what belongs in a safe answer. Remote, hourly, first batch about 10 hours. For CBRN professionals with three or more years in the field, based in the US, UK, Canada, Australia, New Zealand or the EU.","requirements":["Bachelor's degree or higher in a CBRN-adjacent field, or equivalent professional experience","Minimum 3 years of professional experience in at least one CBRN strand (chemical, biological, radiological or nuclear)","Names a primary hazard strand in the expert survey","Based in the US, UK, Canada, Australia, New Zealand or the European Union","English at C1 or above, written and verbal","Passes the expert survey before a bootcamp seat is allocated","Available for the paid bootcamp on Monday and production from Tuesday","Able to commit roughly 10 hours inside the first batch window, closing Sunday","Joins the Sovrano Discord before the bootcamp","Independent contractor, invoices Sovrano directly"],"location":"Remote (United States, United Kingdom, Canada, Australia, New Zealand, or European Union)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["CBRN","Chemical safety","Biosafety","Radiation safety","Nuclear safety","Risk assessment","Hazard communication","Emergency preparedness","AI safety review","Professional English"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":35,"salaryMax":49,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Sovrano AI builds evaluation and training data for frontier AI labs, and this project is about where those models handle hazardous material badly. You take a real user request, look at how the model answered it, and rule on whether the answer was safe, accurate and appropriately framed. Some of what you see will be a model handing out operational detail it should never have given. Just as much of it will be a model refusing a perfectly ordinary question from an emergency planner or a safety officer, and catching that matters as much as catching the dangerous case. The work spans all four hazard classes, and you will be assigned to the strand that matches your background: chemical safety and toxicology, biosafety and biosecurity, radiological protection, or nuclear safety and safeguards. You tell us which in the short survey we send after you apply. This panel is about misuse and safety framing. If your ground is the physics itself, or process engineering, we run separate technical panels you would fit better.","keyResponsibilities":"Assess AI answers on chemical, biological, radiological, nuclear, hazardous-material and emergency-preparedness topics for technical accuracy and safe framing.\nFlag content that gives procedural detail, acquisition routes, weaponization support, unsafe lab guidance or harmful exposure advice.\nFlag the opposite failure too: refusals and hedged non-answers on legitimate professional and public-safety questions.\nWork primarily in your own hazard strand, while holding enough breadth to judge a request that crosses two of them.\nRank competing model answers against each other and say what separates them.\nWrite the reference answer: the response that handles a benign CBRN question properly without crossing into operational detail.\nCatch hallucinated claims, wrong hazard classifications, misstated exposure risks, bad escalation advice and overconfident conclusions.\nApply the project rubric consistently and explain each call in writing.\nEscalate ambiguous or high-risk items for a second review rather than deciding alone.","idealQualifications":"Bachelor's degree or higher in chemistry, toxicology, microbiology, biology, public health, health physics, radiation science, nuclear engineering, environmental health and safety, biosecurity, emergency management, security studies or a related field. Equivalent professional experience also counts.\nThree or more years of professional work in at least one of: chemical safety and toxicology, biosafety and biosecurity, radiation protection and health physics, or nuclear safety, safeguards and nonproliferation.\nDepth in your own strand, plus enough working knowledge of the other three to recognize when a request crosses into them.\nA grasp of hazard categories, exposure pathways, protective measures, decontamination, emergency communication and the public-safety principles around all of it.\nThe judgment to tell an unsafe answer from a merely uncomfortable one, and the confidence to say so in writing.\nEnglish at C1 or above, since the rubric and all written feedback are in English.\nBased in the US, UK, Canada, Australia, New Zealand or the European Union.","niceToHave":"A professional credential in your strand: CIH, CHP, biosafety officer certification, RSO appointment, Chartered Chemist, or equivalent.\nPractical work with GHS classification, biosafety levels, ALARA and source categorization, or IAEA safeguards.\nEmergency response, incident command or exercise experience.\nPrior AI safety, red-teaming, policy review or rubric-based evaluation work.","whatSuccessLooksLike":"You clear the bootcamp on your own merit and start producing on day one of the batch.\nYour judgment on what is safe to answer, and what is not, matches a senior reviewer's without needing a second opinion.\nYour written reasoning tells a reviewer why an answer is wrong, not just that it is.\nYou catch the ordinary professional question the model refused for no good reason, as reliably as you catch the dangerous one it answered.\nYou hold the quality bar across the whole batch, including the last task on Sunday.\nYou are reachable on Discord during the batch and flag ambiguous cases instead of guessing.","contractPaymentTerms":"$35 to $49 per hour, based on your background. Your exact rate is confirmed in writing before you start the bootcamp. Paid hourly for time worked, as an independent contractor.\n\nThe first batch is roughly 10 hours of work. A paid bootcamp runs on Monday, production opens Tuesday, and the batch closes Sunday. There is a good chance the work is extended beyond the first batch, and the people who deliver well on it are the ones we go back to.\n\nPayment runs through Sovrano directly. You invoice us, we do not file or submit tax forms on your behalf.\n\nFully remote, on your own schedule inside the batch window. No fixed shifts and no standing meetings.\n\nHow to apply: apply here with your CV or LinkedIn profile, then complete the short expert survey we send you. Shortlisted applicants get the bootcamp invitation and the project guidelines by email.","createdAt":"2026-09-21 13:47:50.328805+00"},{"id":"52ab2868-3a4d-4fd6-9d31-c404f55f3c5b","title":"Chemical Engineer, AI Model Evaluation","status":"published","description":"Put frontier AI models through real process engineering problems and mark where their reasoning breaks. Remote, hourly, first batch about 10 hours. For chemical engineers with four or more years of hands-on industry work.","requirements":["Bachelor's degree in chemical engineering or closely related, Master's/PhD or PE/Chartered strongly preferred","Minimum 4 years of professional chemical engineering experience, hands-on rather than purely academic","Based in the US, UK, Canada, Australia, New Zealand or the European Union","English at C1 or above, written and verbal","Passes the expert survey before a bootcamp seat is allocated","Available for the paid bootcamp on Monday and production from Tuesday","Able to commit roughly 10 hours inside the first batch window, closing Sunday","Joins the Sovrano Discord before the bootcamp","Independent contractor, invoices Sovrano directly"],"location":"Remote (United States, United Kingdom, Canada, Australia, New Zealand, or European Union)","jobType":"Part-time","domain":"Other Engineering","skills":["Chemical engineering","Process design","Process safety","Thermodynamics","Heat & mass transfer","Separations","Process control","Scale-up","Technical review","Professional English"],"educationLevel":["Bachelors","Masters","PhD","Professional Degree"],"salaryMin":35,"salaryMax":49,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Sovrano AI builds evaluation and training data for frontier AI labs, and this project tests how well those models actually reason about chemical engineering. You give them realistic scenarios, calculations and process problems, then judge what comes back: whether the material and energy balances close, whether the assumptions were stated and defensible, whether the separations logic is feasible, whether the safety reasoning would survive contact with a real plant. Where the model is wrong you write down why, in the terms a colleague would use. Part of the work is also shaping how these models talk about engineering at all, including when they should stop and tell the user to bring in a licensed engineer. The work is remote and hourly, and the first batch runs about ten hours. This is the process engineering panel. If your ground is toxicology, hazard classification and chemical security rather than plant reasoning, apply to CBRN Specialist, AI Safety Review instead.","keyResponsibilities":"Evaluate AI answers across chemical engineering topics for technical accuracy, logical soundness and completeness.\nPush models with realistic scenarios: plant troubleshooting, scale-up questions, sizing logic, process control reasoning.\nMark up model output where assumptions are missing, constraints are ignored, units are inconsistent or the recommendation is unsafe.\nReview and refine prompts, step-by-step solutions and the explanations attached to them.\nCover material and energy balances, phase behavior and thermodynamic property use, unit operations and sizing, distillation and separations feasibility, process control, process safety, and operability at scale.\nSet the standard for how models state assumptions, limitations and professional judgment.\nSay clearly when the right answer is to escalate to a licensed engineer or a site expert.","idealQualifications":"Bachelor's degree in chemical engineering or a closely related field. A Master's or PhD, or PE or Chartered status, is strongly preferred.\nFour or more years of professional chemical engineering experience with real hands-on work: process design, plant and operations engineering, process safety, scale-up, manufacturing, utilities, EHS, quality, R&D, or controls and automation interfaces.\nDeep command of the fundamentals: material and energy balances, thermodynamics, fluid mechanics, heat and mass transfer, reaction engineering, separations and process control.\nComfort with engineering calculations and with defending them, including uncertainty, safety factors and where a method stops applying.\nEnglish at C1 or above, since the rubric and all written feedback are in English.\nBased in the US, UK, Canada, Australia, New Zealand or the European Union.","niceToHave":"Working familiarity with PFDs and P&IDs, Aspen Plus or HYSYS, CHEMCAD, PSM documentation, HAZOP or LOPA.\nPE or Chartered Engineer status.\nPrior AI data training, red-teaming or technical content evaluation.","whatSuccessLooksLike":"You clear the bootcamp on your own merit and start producing on day one of the batch.\nYour judgment on what is safe to answer, and what is not, matches a senior reviewer's without needing a second opinion.\nYour written reasoning tells a reviewer why an answer is wrong, not just that it is.\nYou catch the ordinary professional question the model refused for no good reason, as reliably as you catch the dangerous one it answered.\nYou hold the quality bar across the whole batch, including the last task on Sunday.\nYou are reachable on Discord during the batch and flag ambiguous cases instead of guessing.","contractPaymentTerms":"$35 to $49 per hour, based on your background. Your exact rate is confirmed in writing before you start the bootcamp. Paid hourly for time worked, as an independent contractor.\n\nThe first batch is roughly 10 hours of work. A paid bootcamp runs on Monday, production opens Tuesday, and the batch closes Sunday. There is a good chance the work is extended beyond the first batch, and the people who deliver well on it are the ones we go back to.\n\nPayment runs through Sovrano directly. You invoice us, we do not file or submit tax forms on your behalf.\n\nFully remote, on your own schedule inside the batch window. No fixed shifts and no standing meetings.\n\nHow to apply: apply here with your CV or LinkedIn profile, then complete the short expert survey we send you. Shortlisted applicants get the bootcamp invitation and the project guidelines by email.","createdAt":"2026-09-21 13:47:50.328805+00"},{"id":"1290c181-cbcd-48ba-a149-99593fce06c4","title":"Norwegian AI Response Evaluator (Native Speaker)","status":"closed","description":"Compare pairs of AI-generated Norwegian responses and judge which one is better. For each item you rate both responses on instruction following, truthfulness, coherence and clarity, and length, fact-check every objective claim, and rank the stronger response with your reasoning. The work teaches AI to write and reason better in Norwegian. Native Norwegian is required, though we ask for no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Norwegian speaker","English reading at B2 or above","Reliable internet and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Background in creative writing, journalism, or similar preferred"],"location":"Remote (EU)","jobType":null,"domain":"Language & AI Evaluation","skills":["Norwegian (native)","AI response evaluation","Fact-checking and research","Critical reading","Clear written reasoning","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":12,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models improve when native speakers judge their output carefully. That is the work here.\n\nYou receive items in Norwegian, each with a prompt and two responses generated by our client's language model. For every item you evaluate both responses on instruction following, truthfulness, coherence and clarity, and length. You have search and research tools to fact-check every objective claim, and you document a reliable source for each error you find. Then you write up each response's clearest strengths and areas for improvement, give it an overall quality score, and rank which of the two is preferred, with your reasoning. You work on items in your own field of knowledge, in your native Norwegian.\n\nNo prior AI knowledge is needed. If you are a native Norwegian speaker with solid English and a sharp eye for writing and reasoning, you can do this well. A background in creative writing, journalism, or a similar field is a plus but not required. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Read each prompt and the two AI-generated Norwegian responses, and evaluate each response on instruction following, truthfulness, coherence and clarity, and length\nFact-check every objective claim using the provided search and research tools, and document a reliable source for each error you find\nWrite up each response's clearest strengths and areas for improvement\nGive each response an overall quality score\nRank which of the two responses is preferred and explain your reasoning\nApply reviewer feedback and keep your judgments consistent across the batch","idealQualifications":"Native Norwegian speaker\nEnglish at B2 or above, enough to work from written guidelines\nA sharp eye for writing quality, factual accuracy, and clear reasoning\nComfortable working independently and remotely on task-based work\nReliable internet and a quiet space to focus\nAvailable at least 15 hours per week","niceToHave":"A background in creative writing, journalism, or a similar field\nSubject-matter expertise in a domain you can evaluate confidently\nPrior experience evaluating or reviewing written content, or working from structured guidelines","whatSuccessLooksLike":"Evaluations that hold up when a reviewer checks them against the prompt and responses\nEvery objective claim fact-checked, with a reliable source documented for each error\nClear, specific write-ups of strengths, weaknesses, and the reason for your ranking\nConsistency across the whole batch rather than item-by-item drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-09-09 08:56:50.647451+00"},{"id":"2c6659d9-71af-4705-ba57-f56c8801df51","title":"Danish AI Response Evaluator (Native Speaker)","status":"closed","description":"Compare pairs of AI-generated Danish responses and judge which one is better. For each item you rate both responses on instruction following, truthfulness, coherence and clarity, and length, fact-check every objective claim, and rank the stronger response with your reasoning. The work teaches AI to write and reason better in Danish. Native Danish is required, though we ask for no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Danish speaker","English reading at B2 or above","Reliable internet and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Background in creative writing, journalism, or similar preferred"],"location":"Remote (EU)","jobType":null,"domain":"Language & AI Evaluation","skills":["Danish (native)","AI response evaluation","Fact-checking and research","Critical reading","Clear written reasoning","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":12,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models improve when native speakers judge their output carefully. That is the work here.\n\nYou receive items in Danish, each with a prompt and two responses generated by our client's language model. For every item you evaluate both responses on instruction following, truthfulness, coherence and clarity, and length. You have search and research tools to fact-check every objective claim, and you document a reliable source for each error you find. Then you write up each response's clearest strengths and areas for improvement, give it an overall quality score, and rank which of the two is preferred, with your reasoning. You work on items in your own field of knowledge, in your native Danish.\n\nNo prior AI knowledge is needed. If you are a native Danish speaker with solid English and a sharp eye for writing and reasoning, you can do this well. A background in creative writing, journalism, or a similar field is a plus but not required. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Read each prompt and the two AI-generated Danish responses, and evaluate each response on instruction following, truthfulness, coherence and clarity, and length\nFact-check every objective claim using the provided search and research tools, and document a reliable source for each error you find\nWrite up each response's clearest strengths and areas for improvement\nGive each response an overall quality score\nRank which of the two responses is preferred and explain your reasoning\nApply reviewer feedback and keep your judgments consistent across the batch","idealQualifications":"Native Danish speaker\nEnglish at B2 or above, enough to work from written guidelines\nA sharp eye for writing quality, factual accuracy, and clear reasoning\nComfortable working independently and remotely on task-based work\nReliable internet and a quiet space to focus\nAvailable at least 15 hours per week","niceToHave":"A background in creative writing, journalism, or a similar field\nSubject-matter expertise in a domain you can evaluate confidently\nPrior experience evaluating or reviewing written content, or working from structured guidelines","whatSuccessLooksLike":"Evaluations that hold up when a reviewer checks them against the prompt and responses\nEvery objective claim fact-checked, with a reliable source documented for each error\nClear, specific write-ups of strengths, weaknesses, and the reason for your ranking\nConsistency across the whole batch rather than item-by-item drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-09-09 08:56:47.710527+00"},{"id":"b08b0f96-11d3-42eb-a7b7-9f551db08450","title":"Finnish AI Response Evaluator (Native Speaker)","status":"closed","description":"Compare pairs of AI-generated Finnish responses and judge which one is better. For each item you rate both responses on instruction following, truthfulness, coherence and clarity, and length, fact-check every objective claim, and rank the stronger response with your reasoning. The work teaches AI to write and reason better in Finnish. Native Finnish is required, though we ask for no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Finnish speaker","English reading at B2 or above","Reliable internet and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Background in creative writing, journalism, or similar preferred"],"location":"Remote (EU)","jobType":null,"domain":"Language & AI Evaluation","skills":["Finnish (native)","AI response evaluation","Fact-checking and research","Critical reading","Clear written reasoning","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":12,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models improve when native speakers judge their output carefully. That is the work here.\n\nYou receive items in Finnish, each with a prompt and two responses generated by our client's language model. For every item you evaluate both responses on instruction following, truthfulness, coherence and clarity, and length. You have search and research tools to fact-check every objective claim, and you document a reliable source for each error you find. Then you write up each response's clearest strengths and areas for improvement, give it an overall quality score, and rank which of the two is preferred, with your reasoning. You work on items in your own field of knowledge, in your native Finnish.\n\nNo prior AI knowledge is needed. If you are a native Finnish speaker with solid English and a sharp eye for writing and reasoning, you can do this well. A background in creative writing, journalism, or a similar field is a plus but not required. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Read each prompt and the two AI-generated Finnish responses, and evaluate each response on instruction following, truthfulness, coherence and clarity, and length\nFact-check every objective claim using the provided search and research tools, and document a reliable source for each error you find\nWrite up each response's clearest strengths and areas for improvement\nGive each response an overall quality score\nRank which of the two responses is preferred and explain your reasoning\nApply reviewer feedback and keep your judgments consistent across the batch","idealQualifications":"Native Finnish speaker\nEnglish at B2 or above, enough to work from written guidelines\nA sharp eye for writing quality, factual accuracy, and clear reasoning\nComfortable working independently and remotely on task-based work\nReliable internet and a quiet space to focus\nAvailable at least 15 hours per week","niceToHave":"A background in creative writing, journalism, or a similar field\nSubject-matter expertise in a domain you can evaluate confidently\nPrior experience evaluating or reviewing written content, or working from structured guidelines","whatSuccessLooksLike":"Evaluations that hold up when a reviewer checks them against the prompt and responses\nEvery objective claim fact-checked, with a reliable source documented for each error\nClear, specific write-ups of strengths, weaknesses, and the reason for your ranking\nConsistency across the whole batch rather than item-by-item drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-09-09 08:56:44.927822+00"},{"id":"13e12e9e-455f-46fe-b722-060e7ab7052a","title":"Swedish AI Response Evaluator (Native Speaker)","status":"closed","description":"Compare pairs of AI-generated Swedish responses and judge which one is better. For each item you rate both responses on instruction following, truthfulness, coherence and clarity, and length, fact-check every objective claim, and rank the stronger response with your reasoning. The work teaches AI to write and reason better in Swedish. Native Swedish is required, though we ask for no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Swedish speaker","English reading at B2 or above","Reliable internet and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Background in creative writing, journalism, or similar preferred"],"location":"Remote (EU)","jobType":null,"domain":"Language & AI Evaluation","skills":["Swedish (native)","AI response evaluation","Fact-checking and research","Critical reading","Clear written reasoning","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":12,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models improve when native speakers judge their output carefully. That is the work here.\n\nYou receive items in Swedish, each with a prompt and two responses generated by our client's language model. For every item you evaluate both responses on instruction following, truthfulness, coherence and clarity, and length. You have search and research tools to fact-check every objective claim, and you document a reliable source for each error you find. Then you write up each response's clearest strengths and areas for improvement, give it an overall quality score, and rank which of the two is preferred, with your reasoning. You work on items in your own field of knowledge, in your native Swedish.\n\nNo prior AI knowledge is needed. If you are a native Swedish speaker with solid English and a sharp eye for writing and reasoning, you can do this well. A background in creative writing, journalism, or a similar field is a plus but not required. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Read each prompt and the two AI-generated Swedish responses, and evaluate each response on instruction following, truthfulness, coherence and clarity, and length\nFact-check every objective claim using the provided search and research tools, and document a reliable source for each error you find\nWrite up each response's clearest strengths and areas for improvement\nGive each response an overall quality score\nRank which of the two responses is preferred and explain your reasoning\nApply reviewer feedback and keep your judgments consistent across the batch","idealQualifications":"Native Swedish speaker\nEnglish at B2 or above, enough to work from written guidelines\nA sharp eye for writing quality, factual accuracy, and clear reasoning\nComfortable working independently and remotely on task-based work\nReliable internet and a quiet space to focus\nAvailable at least 15 hours per week","niceToHave":"A background in creative writing, journalism, or a similar field\nSubject-matter expertise in a domain you can evaluate confidently\nPrior experience evaluating or reviewing written content, or working from structured guidelines","whatSuccessLooksLike":"Evaluations that hold up when a reviewer checks them against the prompt and responses\nEvery objective claim fact-checked, with a reliable source documented for each error\nClear, specific write-ups of strengths, weaknesses, and the reason for your ranking\nConsistency across the whole batch rather than item-by-item drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-09-09 08:56:42.049958+00"},{"id":"9682ffc9-0810-459f-818a-1d424cad2554","title":"Dutch AI Response Evaluator (Native Speaker)","status":"closed","description":"Compare pairs of AI-generated Dutch responses and judge which one is better. For each item you rate both responses on instruction following, truthfulness, coherence and clarity, and length, fact-check every objective claim, and rank the stronger response with your reasoning. The work teaches AI to write and reason better in Dutch. Native Dutch is required, though we ask for no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Dutch speaker","English reading at B2 or above","Reliable internet and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Background in creative writing, journalism, or similar preferred"],"location":"Remote (EU)","jobType":null,"domain":"Language & AI Evaluation","skills":["Dutch (native)","AI response evaluation","Fact-checking and research","Critical reading","Clear written reasoning","English (B2+)"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":12,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models improve when native speakers judge their output carefully. That is the work here.\n\nYou receive items in Dutch, each with a prompt and two responses generated by our client's language model. For every item you evaluate both responses on instruction following, truthfulness, coherence and clarity, and length. You have search and research tools to fact-check every objective claim, and you document a reliable source for each error you find. Then you write up each response's clearest strengths and areas for improvement, give it an overall quality score, and rank which of the two is preferred, with your reasoning. You work on items in your own field of knowledge, in your native Dutch.\n\nNo prior AI knowledge is needed. If you are a native Dutch speaker with solid English and a sharp eye for writing and reasoning, you can do this well. A background in creative writing, journalism, or a similar field is a plus but not required. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Read each prompt and the two AI-generated Dutch responses, and evaluate each response on instruction following, truthfulness, coherence and clarity, and length\nFact-check every objective claim using the provided search and research tools, and document a reliable source for each error you find\nWrite up each response's clearest strengths and areas for improvement\nGive each response an overall quality score\nRank which of the two responses is preferred and explain your reasoning\nApply reviewer feedback and keep your judgments consistent across the batch","idealQualifications":"Native Dutch speaker\nEnglish at B2 or above, enough to work from written guidelines\nA sharp eye for writing quality, factual accuracy, and clear reasoning\nComfortable working independently and remotely on task-based work\nReliable internet and a quiet space to focus\nAvailable at least 15 hours per week","niceToHave":"A background in creative writing, journalism, or a similar field\nSubject-matter expertise in a domain you can evaluate confidently\nPrior experience evaluating or reviewing written content, or working from structured guidelines","whatSuccessLooksLike":"Evaluations that hold up when a reviewer checks them against the prompt and responses\nEvery objective claim fact-checked, with a reliable source documented for each error\nClear, specific write-ups of strengths, weaknesses, and the reason for your ranking\nConsistency across the whole batch rather than item-by-item drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-09-09 08:56:38.854698+00"},{"id":"ed25a605-0d44-4568-966a-be724f1024db","title":"Hindi Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Hindi audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Hindi actually sounds across real voices. Native Hindi is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Hindi speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Hindi (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Hindi. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Hindi speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Hindi audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Hindi\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Hindi speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Hindi\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Standard Hindi and regional accents, including the register overlap with Urdu\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"daea27fc-7814-41fb-9e2e-29c30bfee9da","title":"Japanese Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Japanese audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Japanese actually sounds across real voices. Native Japanese is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Japanese speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Japanese (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Japanese. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Japanese speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Japanese audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Japanese\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Japanese speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Japanese\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Standard Tokyo Japanese and regional accents such as Kansai and Tohoku\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"b399bc25-b528-44a4-a439-0f1513db0877","title":"Spanish Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Spanish audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Spanish actually sounds across real voices. Native Spanish is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Spanish speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Spanish (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Spanish. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Spanish speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Spanish audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Spanish\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Spanish speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Spanish\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Peninsular, Rioplatense, Mexican, Caribbean, and Andean Spanish\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"a7fecbac-8cd3-4022-85d6-43c6c3232882","title":"Portuguese Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Portuguese audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Portuguese actually sounds across real voices. Native Portuguese is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Portuguese speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Portuguese (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Portuguese. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Portuguese speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Portuguese audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Portuguese\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Portuguese speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Portuguese\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with European, Brazilian, and African Portuguese (Angolan, Mozambican)\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"7111dade-45f0-443b-8aa2-75be93562d0e","title":"German Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short German audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken German actually sounds across real voices. Native German is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native German speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["German (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":10,"salaryMax":12,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in German. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native German speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe German audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually German\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native German speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken German\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Standard German, Austrian German, and Swiss Standard German\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 10 to 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved. Referral bonuses available.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"6de30886-f83f-45f9-860c-b20237ddabb5","title":"French Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short French audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken French actually sounds across real voices. Native French is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native French speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["French (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in French. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native French speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe French audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually French\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native French speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken French\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Metropolitan, Belgian, Swiss, Quebecois, and West and North African French\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"1891743d-da94-497a-b962-58e0d5a92d0e","title":"Korean Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Korean audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Korean actually sounds across real voices. Native Korean is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Korean speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Korean (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":10,"salaryMax":12,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Korean. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Korean speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Korean audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Korean\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Korean speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Korean\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Standard Seoul Korean and Gyeongsang, Jeolla, and Jeju accents\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 10 to 12 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved. Referral bonuses available.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"13d5b175-9afe-4967-bda3-4fd0c74f945c","title":"Italian Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Italian audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Italian actually sounds across real voices. Native Italian is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Italian speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Italian (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Italian. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Italian speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Italian audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Italian\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Italian speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Italian\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with Standard Italian and Northern, Central, and Southern regional accents\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"0887b7ef-01e5-40c6-b8a0-0759150c637a","title":"Mandarin Chinese Speech Data Specialist (Native Speaker)","status":"closed","description":"Transcribe short Mandarin Chinese audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Mandarin Chinese actually sounds across real voices. Native Mandarin Chinese is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Mandarin Chinese speaker","English reading at B2 or above","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Available at least 15 hours per week","Audio transcription project experience preferred"],"location":"Remote","jobType":"Part-time","domain":"Language & Audio","skills":["Mandarin Chinese (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (B2+)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":8,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Mandarin Chinese. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Mandarin Chinese speaker with solid English and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.\n\nYou should be able to tell Mandarin apart from Cantonese and other Sinitic languages by ear, and flag clips that are not actually Mandarin.","keyResponsibilities":"Transcribe Mandarin Chinese audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Mandarin Chinese\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Mandarin Chinese speaker\nEnglish at B2 or above, enough to work from written guidelines\nA careful ear for accent, dialect, and fluency in spoken Mandarin\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with mainland Putonghua, Taiwanese Guoyu, and regional accents\nExperience with structured labeling or annotation guidelines","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":"Paid at EUR 8 per hour for the pilot, with a minimum of 15 hours per week. Fully remote, task-based, set your own hours. Production rates are still to be confirmed. Paid monthly once your work is approved.","createdAt":"2026-08-28 07:44:05.897058+00"},{"id":"dbf7fe28-2f0b-4170-a85c-5bf694e3ddde","title":"Legal Case Study Author (AI Training)","status":"published","description":"Author regulatory and compliance decision cases grounded in EU law (GDPR, EU AI Act, DORA, MiCA) that train and evaluate frontier AI models. $500 per approved case. Remote.","requirements":["Law degree (LLB, LLM, JD or equivalent) or compliance-focused degree with legal-regulatory work experience","Minimum 2 years of non-academic legal, regulatory, compliance or data-protection experience","Working command of at least two of GDPR, EU AI Act, DORA, MiCA, NIS2, MiFID II","Able to cite regulation at article level","Professional English","Human-authored work only, with no AI use anywhere in the case (file metadata is verified)","Available for a short live walkthrough call during assessment","Remote, working to your own schedule against an agreed submission date"],"location":"Remote (European Union)","jobType":"Contract","domain":"Law","skills":["EU regulatory compliance","GDPR","EU AI Act","DORA","MiCA","Legal drafting","Risk assessment","Executive-level written analysis"],"educationLevel":["Bachelors","Masters","Professional Degree","PhD"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"USD","salaryType":"per_approved_task","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Sovrano AI builds evaluation and training environments for frontier AI models. We are hiring domain experts to author real-world case studies that become the human reference an AI model is measured against. Each case is a paid, self-contained piece of work: you author it, our reviewers assess it, and you are paid on approval.\n\nThis role has immediate paid work: $500 for each approved case study (case brief plus data room). Approved authors will also be among the first experts we reach out to when relevant opportunities arise, with access to future projects available through our expert network.\n\nFor this role the cases are regulatory and compliance decision scenarios under EU frameworks: GDPR, the EU AI Act, DORA, MiCA, NIS2, MiFID II and adjacent regimes. The decision-maker is typically a General Counsel, DPO, Head of Compliance, or an executive committee acting on their advice.","keyResponsibilities":"Author a complete case built around a real decision at a real company, from scoping through submission.\nWrite the case study brief: a firm-and-role persona (never your own name), the audience and the decision in front of it, and a real company (listed EU with public data, or private with disclosed financials) with a one-line justification of why it supports a verifiable decision.\nBuild the thesis as 3 to 5 bullets, each carrying a number, plus a preliminary recommendation with a number, flagged as contingent on the analysis.\nSet out the scoped deliverable plan and the assumptions approach: sources named tier by tier, every assumption labeled sourced or estimated, and a pinned as-of date for the data.\nState the weakest point of your own case, and its mitigant.\nAssemble a slim data room: 10 to 20 distinct, relevant documents behind a single index. Prefer real public documents. Where an internal figure is genuinely undisclosed, construct a clearly labeled case-data extract anchored to a cited public benchmark.\nMake every number in the case trace to a data-room document.\nAnchor each case on a live compliance decision with resources and consequences attached: whether a product launch is compliant and under what conditions, the risk classification of an AI system under the EU AI Act and the obligations that follow, a DORA ICT third-party remediation decision, whether a token or service falls in scope of MiCA and the licensing consequence, or a cross-border data-transfer decision after a regulator ruling. Purely academic or overview-of-the-law topics are rejected.\nCite real regulations at article level (for example Art. 6(1) and Annex III EU AI Act, or Art. 28 GDPR) and real regulator materials: EDPB guidelines, ESMA and EBA statements, published enforcement decisions.\nInclude an in-scope justification line in the brief: why this company and this decision sit inside the named regulatory framework.\nDraw the data room on regulation excerpts, regulator guidance and enforcement decisions, company disclosures and policies, and clearly labeled internal-style extracts where needed.","idealQualifications":"Law degree (LLB, LLM, JD or equivalent), or a compliance-focused degree with demonstrable legal-regulatory work.\n2 to 3 years of non-academic experience in legal practice, regulatory compliance, data protection, or financial-services regulation.\nWorking command of at least two of GDPR, EU AI Act, DORA, MiCA, NIS2 and MiFID II, with the ability to cite them at article level.\nProfessional English, and the ability to write precise, decision-oriented legal analysis for an executive audience.\nFully remote. You work on your own schedule against an agreed submission date.","niceToHave":"In-house experience as, or alongside, a General Counsel, DPO, or Head of Compliance.\nHands-on work with published enforcement decisions or regulator correspondence.\nAdvisory experience on AI systems, ICT third-party risk, or crypto-asset licensing.\nPrior work authoring teaching cases, training material, or AI evaluation data.","whatSuccessLooksLike":"Your case is human-authored end to end, with no AI use anywhere in it. File metadata is verified.\nThe decision at the center of the case is real, in scope, and consequential, not an overview of the topic.\nEvery number is internally consistent and reconciles to a document in the data room.\nYou walk a reviewer through a number and an assumption of our choosing on a short live call, and both hold up.\nThe data room stays slim and relevant: 10 to 20 documents, one index, no filler.\nYour case is approved on the first pass, or after one revision cycle against specific feedback.","contractPaymentTerms":"$500 flat per approved case study (brief plus data room), paid on approval rather than on submission.\n\nApproval follows the Sovrano review gates. If a case needs work you get one substantive revision cycle with specific feedback.\n\nAdditional cases by mutual agreement after your first approval, at the same rate.\n\nFully remote contract work. You set your own schedule against an agreed submission date.\n\nHow to apply: apply here with your CV or LinkedIn profile and 2 to 3 sentences on the decision scenario you would author first, covering the company, the decision, and why you know this ground. Shortlisted applicants receive the full case authoring guidelines and template, and a short intro call before starting.","createdAt":"2026-08-27 13:08:02.700908+00"},{"id":"f4e4856b-d6e4-42c0-a5b7-f1f5f7269ea1","title":"Healthcare Operations Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI handles the business side of healthcare: medical coding, billing, denials, and payer work. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For coding, revenue cycle, CDI, and payer operations professionals in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, or an equivalent coding or HIM credential (CPC, CCS, CCA, RHIA, RHIT, CDIP)","2+ years of hands-on healthcare operations experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Medicine","skills":["Medical coding","Revenue cycle","Denials and appeals","Prior authorization","CDI","Rubric writing"],"educationLevel":["Bachelors","Masters","MBA","Professional Degree","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can explain a coding guideline and still apply it wrongly to a real case. This role puts your experience to work finding out where. You turn cases from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would audit a colleague's work. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own healthcare operations work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"2+ years of hands-on work in medical coding, revenue cycle, CDI, or payer operations preferred\nFluent in the code sets you use every day\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"CPC, CCS, RHIA, or a similar credential (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"a852f07a-96d9-4aec-97e0-2434ba711646","title":"Data Science Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real data science: analysis, experiments, modeling, and SQL. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For data scientists in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher in a quantitative field, completed or in progress","2+ years of applied data science experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Data Analysis","skills":["Python","SQL","Statistics","Experimentation","Machine learning","Rubric writing"],"educationLevel":["Bachelors","Masters","MBA","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models write plausible analysis code quickly, and the conclusions are not always right. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's analysis. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own data science work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"2+ years of applied data science or analytics work preferred\nFluent in Python and/or SQL\nDegree in a quantitative field, completed or in progress\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"A graduate degree in a quantitative field (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"9915a2c9-70ed-4d00-bbdc-3b5e00cdf1cf","title":"Accounting Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real accounting work across audit, tax, close, and reporting. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For accounting professionals in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of hands-on accounting experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Finance","skills":["Technical accounting","Audit","Tax","Month-end close","Reconciliations","Rubric writing"],"educationLevel":["Bachelors","Masters","MBA","Professional Degree","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can describe accounting rules fluently and still reach the wrong conclusion. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's workpapers. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own accounting work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"2+ years of hands-on experience in accounting, audit, or tax preferred\nDegree in accounting, finance, or a related field, completed or in progress\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"CPA, CA, ACCA, or a similar qualification (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"487a2fee-2eb2-4f92-8060-24fbd4163f7f","title":"Finance Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real finance work, from modeling and credit to planning and risk. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For finance professionals in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress","2+ years of hands-on finance experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Finance","skills":["Financial modeling","Valuation","Credit analysis","FP&A","Excel","Rubric writing"],"educationLevel":["Bachelors","Masters","MBA","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can talk about finance with confidence and still get the numbers or the judgment wrong. This role puts your experience to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's analysis. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own finance work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"2+ years of hands-on experience in corporate or investment finance preferred\nDegree in finance, economics, business, or a quantitative field, completed or in progress\nYou build and check financial models yourself\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"CFA, FRM, or securities licenses (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"3f771aea-13b5-4f1d-820c-c1de5b152db2","title":"Life Sciences Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI handles real life sciences work, in three tracks: experimental biology, drug discovery, and computational biology. You build tasks from your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For life scientists in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher in biology or a closely related field, completed or in progress, earned in the US, Canada, Europe, or the UK","2+ years of hands-on lab, drug discovery, or biological data experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Life, Physical & Social Science","skills":["Experimental design","Drug discovery","Bioinformatics","Biological data analysis","Data interpretation","Rubric writing"],"educationLevel":["Bachelors","Masters","PhD"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"Frontier models can explain the biology on paper and still make the wrong call on real data. This role puts your judgment to work finding out where. You turn problems from your own work into realistic tasks, run them through frontier models, and grade the answers the way you would review a colleague's work.\n\nWe are looking for people in three tracks:\n\nExperimental biology. Bench scientists who design and troubleshoot their own experiments. Think of a QC call on a cell-based assay, a flow panel gating review, or reading a Western blot from the raw images.\n\nDrug discovery. Scientists who move programs from target to candidate in pharma, biotech, or an academic discovery unit. Think of a lead optimization decision, triaging a screening campaign, or summarizing a preclinical safety package.\n\nComputational biology and bioinformatics. Analysts who take biological data from raw files to a final interpretation. Think of QC and interpretation of an RNA-seq dataset, a variant prioritization report, or auditing an analysis pipeline.\n\nYou set your own schedule, and clear written reasoning counts as much as scientific depth.","keyResponsibilities":"Build realistic tasks from problems in your own life sciences work\nGive those tasks to frontier AI models and judge the results against the standard of your field\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing","idealQualifications":"2+ years of hands-on experience in a pharma or biotech lab, an academic lab, or a research institute, counted after your undergraduate degree\nDegree in biology or a closely related field, completed or in progress, earned in the US, Canada, Europe, or the UK\nExperimental biology: you design and troubleshoot experiments yourself and can judge raw instrument output, not only run established protocols\nDrug discovery: hands-on discovery or preclinical work; degrees in chemistry, medicinal chemistry, pharmacology, pharmaceutical sciences, or toxicology also count\nComputational biology: you write and can explain your own analysis code in Python or R; biostatistics, computer science, or statistics degrees count when paired with real biological data work\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Programming experience, for example in Python or R\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour tasks come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"0d4cdbfb-4e7c-41fd-a29a-df8ea5626415","title":"Legal Expert, AI Trainer","status":"published","description":"Fully remote, project-based work checking how well frontier AI does real legal work: research, drafting, and review. You write tasks based on your own experience and grade where the model gets it wrong. $40 to $125 USD an hour, at least 10 hours a week, as an independent contractor. For lawyers, senior paralegals, and legal analysts in the US, UK, Ireland, Canada, or Australia.","requirements":["Based in the United States, United Kingdom, Ireland, Canada, or Australia","Bachelor's degree or higher, completed or in progress; law degree preferred","2+ years of legal experience preferred","Available at least 10 hours per week","PayPal account for payment","Willing to work without LLM assistance where the project does not permit it","Use one email address to sign up, take the assessment, and get paid"],"location":"Remote (United States, United Kingdom, Ireland, Canada, or Australia)","jobType":"Part-time","domain":"Law","skills":["Legal research","Contract drafting","Legal writing","Compliance","Due diligence","Rubric writing"],"educationLevel":["Bachelors","Masters","Professional Degree","PhD","MBA"],"salaryMin":40,"salaryMax":125,"salaryCurrency":"USD","salaryType":"hour","weeklyHoursMin":10,"weeklyHoursMax":null,"aboutRole":"AI models can write something that sounds like a lawyer wrote it and still get the law or the reasoning wrong. This role puts your experience to work finding out where. You turn matters from your own work into realistic tasks, run them through frontier models, and judge the answers the way you would review a colleague's draft. You set your own schedule, and the work rewards careful writing as much as expertise.","keyResponsibilities":"Build realistic tasks out of problems from your own legal work, along with the material needed to solve them\nGive those tasks to frontier AI models and judge the results against the standard of your profession\nWeigh paired model answers against each other and explain which one holds up\nWrite grading rubrics for your tasks and set out your reasoning in writing\nReview and improve tasks written by other experts","idealQualifications":"A law degree, or substantial experience as a paralegal or legal analyst\nRegular legal research and drafting in your work\nFull professional or native-level written English\nBased in the United States, United Kingdom, Ireland, Canada, or Australia","niceToHave":"Admission to a bar or law society (welcome, not required)\nHave used Claude or ChatGPT in your professional work","whatSuccessLooksLike":"You clear the skills assessment on your own work, with no AI assistance\nYour task proposals come from real problems in your own practice, and a frontier model genuinely fails them\nYour rubrics are specific enough that another expert would grade the same response the same way\nYou act on reviewer feedback without needing to be walked through it\nYou keep at least 10 hours a week going, week after week","contractPaymentTerms":"$40 to $125 USD an hour. The rate varies with your expertise, education, and experience. Minimum 10 hours a week with no weekly maximum; full-time hours are available.\n\nIndependent contractor, fully remote. Payment runs through PayPal, so you need a PayPal account. As an independent contractor, you handle your own taxes.\n\nGetting on: a skills assessment that takes most people four to six hours. You are paid $240 for it if you pass, and you get one attempt, so take your time and do your best work. Work produced with AI tools is rejected.\n\nAfter you pass: onboarding and a short training project, then you propose a task from your own practice, build it, and run it against frontier models. Each stage is reviewed before you move on, and you are paid for the work you complete.","createdAt":"2026-08-26 19:20:27.96553+00"},{"id":"633a88c6-99c4-4575-a209-f879535f98ce","title":"Mandarin Speech Expert","status":"closed","description":"Review and sharpen how AI writes and reasons in Mandarin. Remote, hourly, flexible. You evaluate AI-generated Chinese content for accuracy, tone, and cultural fit, and write model answers that show the AI what good looks like.","requirements":["Native or near-native Mandarin","English reading/writing at B2 or above","Reliable weekly availability"],"location":"Remote","jobType":null,"domain":"Language & Audio","skills":["Chinese (Mandarin)","English","Localization","Translation","Copyediting","Terminology management","Style guides","Linguistic QA (LQA)","MQM error typology","CAT tools"],"educationLevel":null,"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"This is remote, hourly, flexible evaluation work. You review AI-generated responses in Mandarin, judge how well the model reasons and explains itself, and give expert feedback so its output reads as accurate, natural, and culturally right.\n\nSome days that means rating and comparing several AI answers and explaining which one is better and why. Other days it means writing a strong answer yourself, so the model has a clear example to learn from. You check facts, catch errors in logic or meaning, and make sure terminology, tone, and nuance land correctly for a Chinese audience.\n\nYour calls feed directly into the models built by some of the world's leading AI labs. The judgment you bring is the signal they train on.","keyResponsibilities":"Evaluate and rank AI responses in Mandarin for accuracy, fluency, and how well they fit the context.\nWrite clear model answers and prompts that show the AI what a correct, well-reasoned response looks like.\nFlag errors, bias, and unnatural phrasing, and explain the fix so the model can learn from it.\nCheck that terminology, tone, and cultural nuance are right for a Chinese audience.","idealQualifications":"Native or near-native Mandarin, with a real feel for how the language is used in mainland China.\nComfortable reading and writing English (around B2 or higher) so you can follow prompts and guidelines.\nA sharp eye for detail and good judgment on tone, register, and what reads naturally.\nReliable, and clear in how you document your decisions.\nCurious about AI and happy to work from written guidelines.","niceToHave":"A degree in translation, linguistics, localization, communications, or a related field.\nProfessional translation or localization experience.\nFamiliarity with MQM/LQA quality frameworks (severity, category, root-cause thinking).\nExperience with CAT tools and QA workflows (translation memory, termbases, automated checks).\nPrevious AI data training or evaluation work.","whatSuccessLooksLike":"Your ratings are consistent and easy to trust. Your written examples raise the bar for the model. Your feedback is specific enough that someone could act on it without asking you a follow-up. Over time, the model's Mandarin output gets measurably more accurate and natural because of the calls you made.","contractPaymentTerms":"Hourly contract, paid at €500 per month. Fully remote, set your own hours. Paid monthly once your work is approved. Referral bonuses available.","createdAt":"2026-08-21 14:01:40.485328+00"},{"id":"4e1e5ecd-47d1-4437-bf46-9537cfe62c22","title":"AutoCAD Specialist","status":"closed","description":null,"requirements":["Based in Europe","Please ensure there is no virtual OS installed.","Layouts","Dimensioning","AutoCAD","2D Drafting","CAD Drafting","Layers","Technical drawing","English","Blocks","Screen Recording"],"location":"Europe, Remote","jobType":null,"domain":null,"skills":["Layouts","Dimensioning","AutoCAD","2D Drafting","CAD Drafting","Layers","Technical drawing","English","Blocks","Screen Recording"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD","Professional Degree"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"In this hourly, remote contractor role, you will work as an AutoCAD Specialist (SME) to create realistic, challenging computer-based tasks using AutoCAD and related drafting workflows. You will design practical CAD tasks, document them clearly in SuperAnnotate, ensure your computer setup supports the task, and record yourself executing the task on your computer.\n\nTasks may include 2D drafting, drawing setup, layers, dimensions, blocks, layouts, title blocks, plotting, file organization, annotations, xrefs, or other AutoCAD workflows. You may also be asked to create task prompts, demonstrate correct workflow execution, and ensure that the recording is clear, realistic, and useful for AI training.\n\nYour AutoCAD expertise directly helps improve the world’s premier AI models by teaching them how real users complete practical drafting, documentation, and CAD workflows on a computer.","keyResponsibilities":"Create realistic AutoCAD tasks: Come up with realistic and challenging tasks that can be executed on the computer using AutoCAD, then document the task clearly in SuperAnnotate.\nPrepare the work environment: Ensure your setup supports the task, including having AutoCAD installed, files available, drawings/templates prepared, and any necessary references ready before recording.\nRecord task execution: Record yourself executing the AutoCAD task on your computer while following project instructions and ensuring the workflow is clear, complete, and realistic.\nDocument task context: Provide enough task context, steps, goals, files, and notes so reviewers can understand what the task is and how it was completed.\nDemonstrate practical software use: Complete workflows involving drafting, layers, dimensions, blocks, layouts, plotting, annotations, xrefs, or drawing cleanup where relevant.\nMaintain submission quality: Review your recording and documentation before submission to ensure the task is complete, understandable, realistic, and aligned with project guidelines.","idealQualifications":"Hands-on experience using AutoCAD for drafting, technical drawings, layouts, annotation, dimensions, plotting, or related CAD workflows.\nEnglish proficiency at B2 level or higher to understand project instructions, document tasks clearly, and communicate with the project team.\nComfortable using a computer, managing files, installing/opening required programs, navigating software interfaces, and recording workflows.\nAbility to come up with realistic and challenging AutoCAD tasks that reflect actual drafting, design, documentation, engineering, architecture, or technical drawing use cases.\nStrong understanding of AutoCAD basics such as model space, paper space, layers, lineweights, blocks, dimensions, annotations, layouts, viewports, and plotting.\nAbility to ensure your setup supports the task, including having AutoCAD installed, files available, templates/assets prepared, and any required references ready before recording.\nHigh attention to detail when preparing the task, executing the workflow, and checking that the recording is complete and understandable.\nReliable, self-directed, and able to deliver consistent quality in an hourly, remote contractor workflow across time zones.\nAdditional professional skills are appreciated, including architecture, engineering, drafting, construction documentation, design, sales, administration, coding, operations, or other computer-based workflows.","niceToHave":"Experience with adjacent tools\nFamiliarity with AI tools, though not required","whatSuccessLooksLike":"You will contribute directly to the training data behind real frontier AI systems and built a track record of realistic, high-quality task design that reviewers can trust.","contractPaymentTerms":null,"createdAt":"2026-08-17 16:27:56.472203+00"},{"id":"43a88fe4-c91a-4538-b4ca-1f8264f1cd12","title":"Blender Specialist","status":"closed","description":null,"requirements":["Based in Europe","Please ensure there is no virtual OS installed.","Python","Blender 3D","3D Modelling","English"],"location":"Europe, Remote","jobType":null,"domain":null,"skills":["Python","Blender 3D","3D Modelling"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD","Professional Degree"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"The current generation of AI agents can read a screen, but they still fail badly at doing real professional work inside real desktop software — and Blender is one of the hardest cases there is. Dense modal interfaces, keyboard-driven operators, a 3D viewport where the correct next click depends on the state of the mesh: agents that handle a web form competently fall apart on a modifier stack.\n\nIn this remote, hourly contract role you will help fix that. You will design realistic, challenging Blender tasks, then record yourself completing them on your own computer using recording tools we provide. Each recording captures the full picture of how expert work actually gets done — the sequence of actions, the interface state at every step, and the files before and after — and becomes both a demonstration the model learns from and a benchmark it is measured against.\n\nBecause a model can only learn from work that can be scored, an equally important part of the job is defining what \"done correctly\" means. Alongside each task you will write the objective and the steps you took, and specify how the result should be verified — what the finished scene, mesh, or render has to satisfy for the outcome to count as correct, and which intermediate checkpoints a partially-successful attempt would hit. This is the reward side of the work: in reinforcement learning, an agent improves only when success can be judged consistently, so expert-authored tasks with expert-defined success criteria are the actual bottleneck in training capable agents. Your judgment about what a competent Blender artist would do, and what a plausible-looking-but-wrong result looks like, is the contribution.","keyResponsibilities":"Author production-realistic Blender tasks: Write task briefs that reflect work a studio or client would actually ask for — a low-poly prop to spec, a retopology and UV pass, a shader and lighting setup, a scene assembled to a camera brief — scoped tightly enough that the outcome can be objectively verified.\nRecord expert demonstrations: Set up your own environment for the task (files staged, add-ons and reference assets in place) and record yourself executing it end to end with the provided capture tooling, producing a complete trace of actions, interface state, and file changes.\nDocument the workflow step by step: Log the steps you took in a form another expert could follow and reproduce, in clear professional English.\nDefine verification and success criteria: Specify what the finished result must satisfy, which intermediate milestones mark real progress, and which common wrong paths should not be credited — the rubric a grader uses to score an agent's attempt.\nCover the range of difficulty: Produce tasks from single-tool operations up to long, multi-step workflows that cross applications (reference gathered in a browser, assets from a file manager, delivery into a target format).\nReview and quality-check others' work: For those who move into QA, audit submitted traces for task realism, recording completeness, step-to-action mapping, and whether the stated success criteria genuinely hold for the result produced.\nWork to project quality standards: Keep recordings free of personal information, use only approved applications, and complete all work using your own expertise — the use of closed-source AI assistants is contractually prohibited on this project.","idealQualifications":"2–5 years of professional experience as a Blender-focused 3D artist (contract, studio, or freelance), or equivalent demonstrable practical experience.\nProven proficiency in Blender for production work: modeling and clean topology, modifiers, UV unwrapping, materials and shading, lighting, and rendering in Cycles and/or EEVEE.\nStrong troubleshooting instincts — you can recognize and explain bad topology, shading and normals problems, scale inconsistencies, and non-manifold geometry, not just avoid them.\nA computer you can install our recording tools on and work in daily: Windows 11, Ubuntu (X11), or macOS. Batches are assigned by operating system, and hardware capable of running Blender comfortably alongside a screen recorder is required.\nEnglish proficiency: minimum B2, with the written clarity to document a workflow another expert can reproduce.\nAbility to apply detailed rubrics consistently and make repeatable quality judgments across many similar tasks.\nHighly detail-oriented and self-directed, with strong asynchronous work habits for remote, flexible-hour contract work.\nPrevious experience with AI data training, annotation, evaluation, or benchmark work.","niceToHave":"Experience with other 3D software\nFamiliarity with AI tools, though not required","whatSuccessLooksLike":"You will contribute directly to the training data behind real frontier AI systems and built a track record of realistic, high-quality task design that reviewers can trust.","contractPaymentTerms":null,"createdAt":"2026-08-17 16:22:25.464878+00"},{"id":"6aa508a4-d78a-4e9a-aeb7-7fa8f8299e1e","title":"FreeCAD Specialist","status":"closed","description":null,"requirements":["Based in Europe","Please ensure there is no virtual OS installed.","FreeCAD","Parametric Modeling","CAD","Sketcher","Part Design","Open-Source CAD","Technical drawing","Computer Use","Screen Recording","English"],"location":"Europe, Remote","jobType":"Part-time","domain":"Other Engineering","skills":["FreeCAD","Parametric Modeling","CAD","Sketcher","Part Design","Open-Source CAD","Technical drawing","Computer Use","Screen Recording","English"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"In this hourly, remote contractor role, you will work as a FreeCAD Specialist to create realistic, challenging computer-based tasks using FreeCAD and related parametric CAD workflows. You will design practical FreeCAD tasks, document them clearly in SuperAnnotate, ensure your computer setup supports the task, and record yourself executing the task on your computer.\n\nTasks may include sketching, constraints, Part Design, Part workbench operations, TechDraw, parametric modeling, exports, assemblies or multi-part workflows, technical drawings, or troubleshooting model issues. You may also be asked to create task prompts, demonstrate correct workflow execution, and ensure that the recording is clear, realistic, and useful for AI training.\n\nYour FreeCAD expertise directly helps improve the world’s premier AI models by teaching them how real users complete practical open-source CAD and parametric modeling tasks on a computer.","keyResponsibilities":"Create realistic FreeCAD tasks: Come up with realistic and challenging tasks that can be executed on the computer using FreeCAD, then document the task clearly in SuperAnnotate.\nPrepare the work environment: Ensure your setup supports the task, including having FreeCAD installed, files available, reference images/dimensions prepared, and any required assets ready before recording.\nRecord task execution: Record yourself executing the FreeCAD task on your computer while following project instructions and ensuring the workflow is clear, complete, and realistic.\nDocument task context: Provide enough task context, steps, goals, dimensions, files, and notes so reviewers can understand what the task is and how it was completed.\nDemonstrate practical software use: Complete workflows involving sketches, constraints, parametric features, workbenches, technical drawings, exports, assemblies, or troubleshooting where relevant.\nMaintain submission quality: Review your recording and documentation before submission to ensure the task is complete, understandable, realistic, and aligned with project guidelines.","idealQualifications":"Hands-on experience using FreeCAD for parametric modeling, sketching, constraints, technical drawings, exports, or related CAD workflows.\nEnglish proficiency at B2 level or higher to understand project instructions, document tasks clearly, and communicate with the project team.\nComfortable using a computer, managing files, installing/opening required programs, navigating software interfaces, and recording workflows.\nAbility to come up with realistic and challenging FreeCAD tasks that reflect actual product design, mechanical modeling, technical drawing, or open-source CAD use cases.\nStrong understanding of FreeCAD basics such as Sketcher, constraints, Part Design, Part workbench, bodies, sketches, pads, pockets, booleans, TechDraw, and file exports.\nAbility to ensure your setup supports the task, including having FreeCAD installed, files available, assets prepared, and any required references ready before recording.\nHigh attention to detail when preparing the task, executing the workflow, and checking that the recording is complete and understandable.\nReliable, self-directed, and able to deliver consistent quality in an hourly, remote contractor workflow across time zones.\nAdditional professional skills are appreciated, including mechanical design, product design, engineering, technical documentation, coding, sales, administration, operations, or other computer-based workflows.","niceToHave":"Experience with other CAD packages\nFamiliarity with AI tools, though not required","whatSuccessLooksLike":"You will contribute directly to the training data behind real frontier AI systems and built a track record of realistic, high-quality task design that reviewers can trust.","contractPaymentTerms":null,"createdAt":"2026-08-17 16:13:32.423134+00"},{"id":"9b663ff9-6d18-4b03-95da-26395e9e6adf","title":"Robotics Motion Capture","status":"closed","description":"Review and refine written descriptions of human and robotic motion from motion-capture footage. You correct AI-drafted captions into accurate, model-ready ground truth. Remote, flexible hours, paid per task.","requirements":["Fluent written English","Reliable internet and a computer","Able to commit consistent weekly hours"],"location":"Remote","jobType":null,"domain":"Miscellaneous","skills":["Written English","Attention to detail","Motion description","Data annotation","AI evaluation","Consistency","Remote self-management"],"educationLevel":["High School"],"salaryMin":0.6,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"per_approved_task","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Project Marrow is one of Sovrano AI's robotics efforts, run with leading AI labs. The goal is to teach AI systems to understand movement by giving them precise, human-written descriptions of what happens in motion-capture clips.\n\nYou are the human quality layer. Each clip arrives with an AI-generated first-draft caption. You watch the clip, correct what's wrong, sharpen what's vague, and confirm what's right. The description you finalize becomes the training signal the model learns from, so accuracy and consistency are the whole job.\n\nNo prior experience is required, though it is valued. We onboard you and give you a clear house style to work from.","keyResponsibilities":" Watch short motion-capture clips and describe the motion in clear, plain, present-tense language\nReview the AI-generated draft for each clip and correct inaccuracies\nState what moves, in what direction, and roughly how far, and whether the subject stays in place or travels\nApply the project's house style consistently across every clip\nHedge anything you cannot be certain of, and flag clips that are genuinely faulty\nMaintain a steady pace without letting quality slip\nAct on reviewer feedback to keep improving","idealQualifications":"Strong written English and an eye for precise, unfussy description\nCareful and detail-oriented, small differences matter here\nReliable and consistent, you keep a steady pace and your quality holds\nAble to learn a house style quickly and apply it the same way every time\nComfortable working independently in a remote, self-directed setup","niceToHave":"Background in linguistics, kinesiology, animation, sports, dance, or physical therapy\nPrior experience with data annotation, labeling, or AI evaluation work\nFamiliarity with motion capture or 3D animation","whatSuccessLooksLike":"Your captions are accurate, consistent, and need little rework. You keep a dependable volume, your reject rate stays low, and reviewers trust your judgment. Strong, consistent evaluators move into higher-volume work and reviewer roles as the project grows.","contractPaymentTerms":"Compensation is €0.70 per approved task, with performance bonuses possible. Flexible hours. You claim clips, complete them, and submit for review; approved work is paid out on the standard cycle.","createdAt":"2026-08-13 14:27:43.563454+00"},{"id":"c920900a-7541-48bf-a56a-c097937a580f","title":"Robotics Data Annotator","status":"published","description":"Pre-qualify now for paid robotics annotation work on upcoming Sovrano projects. The qualification is free, fully remote, and takes about two hours. When you are placed on a project, the rate is $5 to $15 per hour depending on experience and assessed accuracy.","requirements":["Read the step-by-step guide before you begin.","Working English, reading and writing","A laptop or desktop computer, not a phone","A reliable internet connection","Willingness to complete a free qualification of about two hours","Sign up on our partner platform using the same email address you apply with","Upload a screenshot of your completed exam screen","Able to set up two factor authentication on your phone"],"location":"Remote","jobType":"Part-time","domain":"Miscellaneous","skills":["Data annotation","Video annotation","Attention to detail","Written English","Following written guidelines","Independent remote work"],"educationLevel":["High School","Bachelors"],"salaryMin":5,"salaryMax":15,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":null,"weeklyHoursMax":40,"aboutRole":"Before you start, read the step-by-step guide. It walks through the whole process with screenshots, including exactly what your final screenshot needs to show. An unclear or incomplete final screenshot is the one thing that will stop your application, so it is worth ten minutes before you begin.\n\nSovrano places annotators on robotics data projects for AI companies building the next generation of robot learning models. The work is describing what a robot or a human hand is doing in short video clips, frame by frame or as a caption. Things like twisting left, loosening a bolt, gripping a handle. Precision and consistency matter far more than speed or creativity.\n\nImportant: There is no immediate project for this role; however, if qualified, you will be among the first experts we reach out to when relevant opportunities arise. This will also provide you with access to future projects available through our expert network.\n\nYou complete a free qualification now, on our partner's annotation platform. Projects open with little notice, which is why we qualify people in advance rather than scrambling when one lands.\n\nThe qualification is an English check, a typing check, optional practice tasks, and a 5-task exam. It is free, you do it in your own time, and nobody pays anything at any point.\n\nOne rule matters more than the rest: sign up on the partner platform using the same email address you apply with here. That is how we confirm your result. A different address means we cannot verify that you passed.\n\nFinishing the exam is not the last step. Once it is done, come back and complete the short completion form we link in your application email. It asks for the email address on your platform account, the referral code you used, and a screenshot of your completed exam. We cannot verify you without it, so your application is only complete once that form is in.","keyResponsibilities":"Describe robotic and manual actions in short video clips, frame by frame or by caption\nApply the project label guide consistently across every task, not just the first ten\nKeep accuracy high across a full working session\nFlag unclear, corrupted or low quality clips instead of guessing\nMeet the agreed turnaround once you are allocated to a project\nRaise questions early when a clip does not fit the guide","idealQualifications":"Working English, both reading and writing\nComfortable following a written label guide precisely and repeatedly\nA laptop or desktop computer and a reliable internet connection\nAble to work independently, remotely, without supervision\nAble to set up two factor authentication on your phone\nAvailable for at least a few hours a week when a project opens","niceToHave":"Previous annotation, labelling or data QA experience\nExposure to robotics, video annotation or computer vision datasets\nFast and accurate typing\nAvailability at short notice when a project opens\nExperience working to a written spec in a remote team","whatSuccessLooksLike":"You pass the qualifying exam and enter our qualified pool\nYour annotations need little or no rework when reviewed\nYou stay consistent across a full session, not just the first few tasks\nYou raise ambiguity instead of guessing\nYou are someone we can place on the next project without retraining","contractPaymentTerms":"Freelance, project based, fully remote.\n\nThe qualification is free and unpaid. When you are allocated to a project, you are paid hourly at a rate between $5 and $15 per hour, set by your experience and your assessed accuracy, and confirmed to you in writing before any work starts. Payment is monthly against approved hours.\n\nHours vary by project and are agreed with you before you begin. This posting is a pre-qualification for upcoming projects. It does not guarantee an offer of work or any specific number of hours.","createdAt":"2026-08-04 14:07:52.774721+00"},{"id":"feac9d8e-187e-441c-a9b5-15d00fada880","title":"Catalan Speech Expert (Native Speaker)","status":"closed","description":"Transcribe short Catalan audio clips and label who is speaking. For each clip you write out what is said and record the number of speakers, the gender of each, and how fluent they sound. The work trains AI on how spoken Catalan actually sounds across real voices. Native Catalan is required, though we require no specific degree or background. Fully remote, task-based, work at your own pace.","requirements":["Native Catalan speaker","English at C1 to C2","Reliable internet, headphones, and a quiet workspace","Comfortable with remote, task-based work and following written guidelines","Audio transcription project experience preferred"],"location":"Remote","jobType":"Internship","domain":"Language & Audio","skills":["Catalan (native)","Audio transcription","Active listening","Speaker identification","Attention to detail","English (C1-C2)"],"educationLevel":["High School","Bachelors","Masters","MBA","EMBA","PhD","Apprenticeship","Professional Degree"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":15,"weeklyHoursMax":null,"aboutRole":"AI models learn spoken language from carefully transcribed and labeled audio. That is the work here.\n\nYou receive batches of short audio clips in Catalan. For each one you produce an accurate transcript of what is said, then answer a small set of structured questions about the voices in the clip: how many people are speaking, the gender of each speaker, and how fluent each sounds (native versus non-native). The value is your ear as a native speaker, so the labels have to reflect what you actually hear, not a guess. Clear audio takes only a few minutes per clip; harder clips with overlapping or background speech take longer.\n\nNo prior AI knowledge is needed, and there is no education requirement. If you are a native Catalan speaker with solid English (C1/C2) and a good ear, you can do this well. Requires a minimum of 15 hours of work per week.","keyResponsibilities":"Transcribe Catalan audio clips accurately, following the formatting and spelling conventions in the project guidelines\nFor each clip, record the number of distinct speakers\nFor each speaker, label gender and fluency (native or non-native)\nFlag clips where the audio is unclear, corrupted, or not actually Catalan\nKeep your transcripts and labels consistent across the full batch\nApply reviewer feedback and correct your work where the guidelines call for it","idealQualifications":"Native Catalan speaker\nEnglish at C1 to C2\nA careful ear for accent, dialect, and fluency in spoken Catalan\nComfortable working independently and remotely on task-based work\nReliable internet, good headphones, and a quiet space to listen\nAvailable at least 15 hours per week","niceToHave":"Prior experience on an audio transcription or subtitling project\nFamiliarity with regional Catalan varieties (Central, Valencian, Balearic, Northern)\nExperience with structured labeling or annotation guidelines.","whatSuccessLooksLike":"Transcripts that match the audio word for word, with the project's spelling and formatting conventions applied consistently\nSpeaker counts and gender and fluency labels that hold up when a reviewer checks them against the clip\nConsistency across the whole batch rather than clip-by-clip drift\nSteady, reliable throughput once you are up to speed","contractPaymentTerms":null,"createdAt":"2026-07-22 15:41:54.241243+00"},{"id":"f9586d6d-89e4-4ba2-b51a-9268a5a308ac","title":"Financial Modeling Expert (IB / PE / Consulting)","status":"published","description":"Build financial models and investor decks that train frontier AI on how finance actually works. You propose a realistic mandate, get a short case brief approved, then deliver the data room, a full Excel model, and a deck or memo, reviewed by a senior finance professional. Remote and project-based, $1,000 per approved task. For investment banking, private equity, and MBB professionals based in the US or Europe.","requirements":["4+ years in investment banking, private equity, or corporate development, VP or senior associate level, at a recognized institution","Or ex-MBB engagement manager with hands-on transaction modeling experience","Has personally built LBO, DCF, M&A, three-statement, and valuation models from scratch on live deals","Domiciled in the US or Europe","Builds the brief, model, and deck themselves; AI tools for background research only","Willing to share past work samples or case studies as part of screening"],"location":"Remote (US or Europe)","jobType":"Part-time","domain":"Finance","skills":["Financial modeling","LBO modeling","DCF / valuation","M&A modeling","Three-statement modeling","PowerPoint / investor decks","Financial analysis","Sector research","Data room assembly"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD"],"salaryMin":1000,"salaryMax":null,"salaryCurrency":"USD","salaryType":"per_approved_task","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"AI labs need finance work produced to a professional standard so their models can learn from genuine expert reasoning. Your judgment as a finance professional is the product.\n\nYou produce institutional-grade casework on a realistic, forward-looking mandate you propose yourself, for example a take-private of a German industrial by a UK sponsor, or a sum-of-the-parts review of a listed European conglomerate. Each task has two steps.\n\nStep 1, the case brief. You submit a short brief: the company, the persona and audience, the thesis, a numeric recommendation, the model scope tab by tab, and your sourcing approach. An automated check and a senior reviewer approve it before you build anything. Most briefs take one round of feedback.\n\nStep 2, the deliverable. Once the brief is approved you build and submit three things: a data room of the public source documents you used (filings, presentations, research, press), a full Excel model with a Source Log tab and a short Verifier sheet that names your key inputs and outputs, and a presentation document (deck or memo). The model and deck go through an automated quality audit, then a senior reviewer checks them against the acceptance criteria. On approval you are paid in full.\n\nSix case types:\n\n1. Pitch, Buyside Opportunity (M&A): banker-style acquisition pitch with target overview, strategic rationale, valuation (DCF plus trading and transaction comps), synergies, accretion/dilution, and financing.\n\n2. Pitch, Strategic Options / SOTP: sum-of-the-parts review of a multi-segment business, segment-level modeling and valuation, break-up vs. hold, divestiture or carve-out scenarios.\n\n3. Client, Sellside Materials (Teaser / CIM): marketing documents for a sale process, from a short teaser to a full information memorandum covering business, market, financials, and outlook.\n\n4. Fairness Opinion: board-ready valuation triangulating DCF, comps, precedents, and an LBO check, with a football field and sensitivity tables.\n\n5. Market Report: an industry study covering market sizing, growth drivers, competitive landscape, value chain, M&A activity, and outlook.\n\n6. Preliminary Investment Committee Memo: a buy-side internal memo with investment thesis, business quality, returns math (entry/exit, IRR bridge), key risks and mitigants, and a diligence plan.\n\nChoosing a case. Pick a company you know well, ideally US or European, and a deal that a real deal team could plausibly be working on today, with a named buyer, seller, board, or investment committee as the audience. Briefs on companies we already cover are rejected, so check with us before you start. Avoid banks and insurers (other financial institutions and fintech are fine), real estate, and oil and gas. Power and utilities, mining, and infrastructure are all fine.\n\nAI use. AI tools may be used for background research only. The brief, the model, and the deck must be built by you. Every submission is audited automatically and reviewed by a senior finance professional. AI-generated deliverables are rejected and end the engagement, unpaid.","keyResponsibilities":"Propose a realistic mandate and write a case brief that a senior reviewer can approve\nBuild the complete financial model from scratch (three-statement, DCF, LBO, M&A, comps as the case requires), with every assumption sourced and logged\nAssemble the data room of public documents your model and deck rely on\nProduce the deck or memo to investment-committee standard, with every number reconciling to the model\nRespond to reviewer feedback and revise until the deliverable meets the acceptance criteria\nWork only from public or illustrative information, never confidential or client-identifying data","idealQualifications":"4+ years in investment banking, private equity, or corporate development, VP or senior associate level, at a recognized institution\nOr ex-MBB engagement manager with hands-on transaction modeling experience\nHave personally built models from scratch on live deals: LBO, DCF, M&A, three-statement, valuation\nComfortable producing investor-grade PowerPoint alongside the numbers\nDeep working knowledge of US or European capital markets","niceToHave":"Sector depth in industrials, TMT, healthcare, consumer, power and utilities, or infrastructure\nCross-border transaction experience\nA portfolio of past work samples or case studies you can share","whatSuccessLooksLike":"Case briefs approved in one round\nDeliverables that pass review on the first or second pass\nModels that are correct, clean, fully sourced, and defensible under scrutiny\nDecks that tell a story an investment committee would act on\nSteady throughput: your first task takes around 15-20 hours, later ones 10-15","contractPaymentTerms":"$1,000 per approved task. Paid on approval, no hourly billing.\nReviewer feedback within 2 working days of each submission. Up to 5 resubmissions per task.\nProject-based and fully remote. Work on your own schedule. Most contributors deliver 2-3 tasks a month alongside a full-time role.","createdAt":"2026-07-20 10:06:20.57424+00"},{"id":"53f134c0-6393-4143-8020-edd9ae941a99","title":"Sovrano AI Summer Internship - Part Time","status":"closed","description":"Join Sovrano AI as a Summer Intern and help teach today's frontier AI models how skilled people actually get work done. Sovrano AI connects university students with AI labs building advanced AI systems, giving interns the opportunity to support real client projects across multiple verticals.\n\nIn this role, you will complete real professional tasks on your own computer and show the AI how you do them. Think of the everyday work you would do in an internship or a first job: building a presentation on a given topic, putting together a document, working through a spreadsheet, researching something online, or moving through the steps of a multi-part task. You do the work the way a capable professional would, and that becomes the material the AI learns from. Because Sovrano works with clients across different AI labs and project types, your assignments will vary throughout the internship, but the core focus stays the same: showing AI systems how humans carry out real work, so they can do it more reliably and accurately.\n\nNo prior AI knowledge is required. You will learn on the job and receive access to the Sovrano AI Curriculum, which includes structured training on how large language models work, how AI systems learn from people, and what good work looks like across business and cross-lingual tasks.\n\nThis is a 12-week, part-time internship requiring approximately 20 hours per week. The role is fully remote and includes a compensation of 500 EUR per month.","requirements":["Currently enrolled in a university program","Available 20 hours per week for the full 12-week program","Fluent written and spoken English (any additional language is a plus)","Strong attention to detail and good reasoning skills"],"location":"Remote","jobType":"part-time","domain":"Miscellaneous","skills":["Critical thinking","Attention to detail","Written communication","Analytical reasoning","Following detailed guidelines","Cross-lingual evaluation (a plus)"],"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"Sovrano AI connects university students with the AI labs building today's frontier models. As an intern, you'll show these systems how real work gets done, completing everyday professional tasks on your own computer so the AI can learn from how a capable person actually works. You'll work directly with our AI lab clients across a range of verticals. The work sits at the center of how modern AI gets built, and you'll be matched to client projects based on where your judgment is most useful, so the day-to-day stays varied and you see how your work feeds into a real development pipeline. No prior AI knowledge required. You'll learn everything on the job, backed by full access to the Sovrano AI Curriculum.","keyResponsibilities":"Complete real professional tasks on your own computer, from building presentations to working through documents and spreadsheets, the way a capable professional would\nDemonstrate how you work through multi-step tasks so AI systems can learn from your process, not only the finished result\nProvide clear, structured feedback that shapes how models learn to carry out real work\nApply client-specific standards for quality and accuracy across business and cross-lingual tasks\nFlag edge cases, inconsistencies, and quality issues for reviewer escalation\nAdapt to new project types as you're matched to different AI lab clients","idealQualifications":"Currently enrolled in or recently graduated from a university program (any field)\nStrong reading comprehension and clear, precise judgment\nComfortable following detailed guidelines and applying them consistently\nAble to work remotely\nCurious about AI and motivated to learn how frontier models are built","niceToHave":"Fluency in any language other than English \nExperience in a specialized domain (law, medicine, finance, sciences, humanities, etc.)\nPrior experience with detailed analytical or research work\nFamiliarity with AI tools, though not required","whatSuccessLooksLike":"By the end of the program, you'll have contributed directly to the evaluation work behind real frontier AI systems, built a track record of consistent, high-quality judgment across client projects, and completed the Sovrano AI Curriculum to become genuinely AI-literate. You'll understand how modern AI gets built from the inside, a credential most people never get access to.","contractPaymentTerms":"- 12-week summer program\n- Part-time, 20 hours per week\n- Fully remote\n- Compensation: 500 EUR per month\n- Includes full access to the Sovrano AI Curriculum","createdAt":"2026-06-18 09:33:05.3092+00"},{"id":"2be103f5-faa6-4017-97f3-a3566f727012","title":"BARCELONA: Robotics - First-Person Data Collection","status":"closed","description":"Sovrano AI is running a paid data-collection project in Barcelona. We’re recording short, first-person (“egocentric”) video of people doing everyday tasks  (food prepping, knitting, stitching, fixing a bike, cleaning your home) so the footage can be used to train the next generation of household robots.","requirements":["Natural pace. Don’t slow down to “show” the camera; don’t speed up either. Act the way you normally would.","No unnecessary walking. Unless the task requires it, stay put.","Multi-step tasks only. A real task with multiple stages, not a single repetitive motion."],"location":"Barcelona","jobType":"contract","domain":"Miscellaneous","skills":null,"educationLevel":["Bachelors","Masters","MBA","EMBA","PhD"],"salaryMin":10,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"hour","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"You wear a small headset camera while doing everyday tasks. That’s it. You’re not performing for the camera; you’re doing exactly what you’d be doing anyway.\nWe pay per hour of recorded footage that meets the technical spec.","keyResponsibilities":"Put on the headset camera (we provide it; lightweight, no setup on your end)\nDo your normal tasks (cook, sew, repair, build, clean, assemble, or even grabbing a coffee or walking the dog) for up to 5 hours at a time (SD card memory limit)\nGo about it at your natural pace; don’t slow down or speed up for the camera\nHand the headset back at the end of your session","idealQualifications":"No prior experience with recording or AI is needed. If you can cook, clean, grab a coffee, walk a dog, play an instrument, ... you can do this.","niceToHave":null,"whatSuccessLooksLike":null,"contractPaymentTerms":"Hourly rate of €10/hr, paid per session.\nPayment is processed after review of your footage.","createdAt":"2026-05-12 15:30:52.998492+00"},{"id":"0eecadb3-f226-4000-be17-1da24942d5e6","title":"AI Evaluation Intern - STEM & Engineering","status":"closed","description":"Sovrano AI connects university students with AI labs building frontier models. Students do the human evaluation work that makes AI systems safer and more accurate: RLHF, data annotation, response ranking. 12-week, full-time (35-40 hrs/week), fully remote from Europe. ~20% structured AI training, ~80% hands-on evaluation. No prior AI knowledge required.","requirements":["Currently enrolled in a Bachelor's or Master's in Engineering (any field), Computer Science, Mathematics, Physics, or a closely related STEM program","Fluent in English plus at least one other European language","Strong quantitative skills (you should be comfortable with graduate-level math)","Comfortable working independently in a fully remote setup","Reliable internet connection and a quiet workspace"],"location":"Remote (Europe)","jobType":"Internship","domain":"Other Engineering","skills":["Mathematical Reasoning","Code Review","Technical Writing","Python","RLHF"],"educationLevel":["Master's","Bachelor's"],"salaryMin":500,"salaryMax":null,"salaryCurrency":"EUR","salaryType":"month","weeklyHoursMin":null,"weeklyHoursMax":null,"aboutRole":"We're hiring engineering and STEM master's students to evaluate AI-generated technical and mathematical content. This is a 12-week, full-immersion internship. The work is technical, hands-on, and paid.\nAround 20% of your time goes toward structured AI training: how large language models work, how RLHF fits into the development pipeline, what good evaluation looks like and why it matters. No prior AI knowledge is required. You'll learn everything on the job.\nThe other 80% is live evaluation work on production AI models. This is the core of the internship. Your quantitative skills and technical reasoning are the reason you're here.","keyResponsibilities":"1. Evaluating AI on mathematical and quantitative reasoning. You'll get two AI-generated solutions to the same math or engineering problem and decide which one is better. These cover calculus, linear algebra, probability, statistics, and optimization at the undergraduate and early graduate level. You're not just checking whether the final answer is right. You're tracing the reasoning chain step by step, identifying where errors creep in, and ranking explanatory clarity. AI models are getting better at math, but they still make subtle mistakes that only someone with genuine quantitative training can spot. This is called RLHF (reinforcement learning from human feedback), and it's the single highest-demand category in AI evaluation right now.\n2. Evaluating AI-generated code. AI coding assistants are used by over 50% of professional developers. AI labs need evaluators who can review AI-generated code across languages (Python, JavaScript, SQL, C++, and others) and assess it for correctness, efficiency, readability, and whether it actually solves the stated problem. You'll compare multiple AI-generated solutions and rank them. Competitive programming benchmarks and real-world software engineering tasks are both in scope.\n3. Fact-checking AI in your native language. AI models hallucinate. They produce plausible-looking but wrong calculations, cite formulas incorrectly, and confidently misapply technical concepts. You'll catch them. If you speak German or French alongside English, you'll review technical AI outputs in both languages and flag errors. Most AI evaluation today only covers English. Technical content in European languages is almost entirely unevaluated.\n4. Red-teaming AI for safety. The EU AI Act now requires companies deploying AI to test their models for harmful outputs. You'll try to break them in your domain. That means finding prompts that lead to dangerously incorrect technical calculations, flawed engineering recommendations, or outputs that could cause real-world harm if someone trusted them. Documenting failure modes in technical domains is one of the highest-value evaluation tasks.\n5. Contributing to industry benchmarks. The evaluation work you do here feeds directly into the benchmarks top-tier AI labs use to assess and publicly report the factual capabilities of their models. When OpenAI, Google DeepMind, or Anthropic measure how well their models reason about STEM subjects, datasets built by evaluators like you are part of what they test against.","idealQualifications":"Currently enrolled in a Master's in Engineering (any field), Computer Science, Mathematics, Physics, or a closely related STEM program\nFluent in English plus at least one other European language\nStrong quantitative skills (you should be comfortable with graduate-level math)\nComfortable working independently in a fully remote setup\nReliable internet connection and a quiet workspace","niceToHave":"Prior experience in engineering, software development, data science, or technical R&D\nCompetitive programming, research publications, or applied math experience\nExperience working across cultures or in international teams","whatSuccessLooksLike":null,"contractPaymentTerms":"Depending on experience (paid)","createdAt":"2026-04-17 00:22:31.581602+00"}]}}