Help shape standards for nuclear and radiological safety evaluations used in advanced-model training. You will set evaluation criteria, guide safe handling of sensitive technical information, and define when security concerns should be escalated. Your subject-matter expertise matters more than prior AI experience.
Help assess how frontier AI models handle difficult, real-world questions in your area of expertise. You’ll shape technical scenarios and problem sets, write reference answers and grading criteria, then judge model responses for accuracy, depth, and sound domain reasoning. The work centers on research and evaluation rather than production engineering, across seven specialist tracks.
A finance specialist will help a leading AI lab assess and improve how its models handle real financial work. You will review finance tasks and model responses, write task instructions and reference solutions, and shape benchmarks with the research team. The full-time W-2 role requires 40 hours per week for an initial six-month engagement and hybrid work in the Bay Area, with on-site attendance multiple days each week when required.
Work on an AI safety initiative by testing models for ethical jailbreaks, prompt injection, and tool-use abuse. You’ll develop adversarial evaluation methods, probe multi-turn weaknesses, maintain regression tests, and share findings with technical and non-technical audiences. AI experience is not required; relevant domain knowledge matters.
Applied economists, econometricians, and quantitative social scientists create empirical economics problems based on their professional work. You will develop examples involving causal questions, panel data, choice models, or forecasting, then produce reference answers with clear reasoning. Frontier AI labs use this work to assess model judgment. The independent contractor role runs on an async schedule, requires at least 10 hours per week, and has no weekly maximum.
Mercor keeps this listing open to build a pool of biotechnology experts for future AI training and evaluation projects. You may write development problems and reference answers, assess model-generated process and assay work, and identify proposals that could fail on developability or regulatory grounds. Projects may start with short notice, and matching can take anywhere from a week to several months.
Apply expertise in warfighting, weapons targeting, defense operations, and international humanitarian law to shape AI training and evaluation for defense and humanitarian scenarios. You’ll develop frameworks for assessing military and civilian uses of technology, prohibited assistance, and conflict-related risks. No prior AI experience is required; the role calls for domain knowledge and clear analysis for interdisciplinary audiences.
AfterQuery engages certified court interpreters to create and assess legal interpretation scenarios used to evaluate AI models. You will draw on courtroom and legal experience to write reference interpretations, explain terminology and register choices, and judge whether model responses preserve the intended legal meaning in both working languages. The work is remote, asynchronous contract work with a minimum commitment of 10 hours per week.
Bring hands-on business or sales operations expertise to work with a leading AI lab’s research and program management teams. You’ll assess operational tasks and model responses, write guidance and reference solutions, and help shape evaluations that measure model performance. Cincinnatus LLC employs this full-time W-2 role, with placement on a client team. The hybrid position requires Bay Area residence and on-site work multiple days each week when required.
Use your military operations and DoD reporting knowledge to review reporting content and help train AI systems. You’ll assess accuracy, clarity, terminology, and reporting standards, then give written feedback and shape realistic training examples. Prior AI experience is not needed. The work is remote and supports a customer project.
AfterQuery engages project managers as contractors to create and assess project delivery scenarios for frontier AI evaluations. You’ll write scenarios that cover kickoff through retrospectives, explain the reasoning behind reference decisions, and judge model responses for realism and stakeholder awareness. The work is remote and asynchronous, with a minimum commitment of 10 hours per week and no set maximum.
Assess how well AI systems reason through advanced security problems, from vulnerability analysis and exploit development to secure software engineering. You’ll judge technical reports for accuracy, exploitability, root-cause analysis, and mitigation quality, then help build benchmarks for AI security capabilities. The role suits researchers with a strong offensive security record and recognized technical contributions.
Create realistic Department of Defense staff-writing examples and evaluation materials for an AI training project. You will develop prompts, source artifacts, model answers, and scoring guidance for formats such as WARs, SITREPs, executive summaries, and performance reports, then refine them through project feedback. Applicants must be based in the United States; prior AI experience is not needed.
Create and assess health information management scenarios that test how AI models handle privacy, release of information, and record integrity. You’ll write reference answers grounded in regulatory reasoning and review model responses for compliance risks. The work is remote and asynchronous, with ongoing contractor projects and a minimum commitment of 10 hours per week.
Mercor keeps a standing pool of biologists for short-notice AI training and evaluation projects. You may write research problems with reference answers, assess model-generated experimental designs and interpretations, and identify answers that fail to account for real bench conditions. Apply with a resume, confirm where you work from, and complete an AI interview of about 20 minutes. Project invitations may arrive within a week or take several months.
Use your military writing and reporting background to create and assess government-style material for AI training. You will work from prompts and supplied artifacts to produce reports, summaries, updates, model examples, and grading guidance. The project also calls for feedback on AI-generated reports and insight into Department of Defense documentation practices. No prior AI experience is needed, and contributors must be based in the US.
Create and validate research-grade benchmark tasks that test frontier AI models across machine learning. The role suits researchers with a master's or PhD, at least one first-author paper, and hands-on experience writing ML code for training, evaluation, or systems work. Work remotely on an asynchronous, flexible schedule of 5 to 40 hours per week, with hourly compensation of $150.
Prepare French speech recordings for speech and text-to-speech systems by editing audio, checking technical quality, and assessing pronunciation and naturalness. The work covers large datasets and requires native or near-native French fluency, strong listening skills, and experience with speech recordings. You will help ensure files meet the specifications used in machine learning training workflows.
Bring naval or maritime expertise to a project supporting Navy decision-making at the Pentagon. You will analyze operational, programmatic, or intelligence information, develop recommendations, and brief senior audiences while contributing real-world expertise to the training of AI systems. The work takes place on-site in Arlington, Virginia, or the DC area. Previous AI experience is not required.
Create and assess process design and safety scenarios that AI labs use to evaluate model judgment. You’ll write reference answers, review model output, and catch designs that could fail on safety or feasibility at scale. The role suits practicing chemical engineers with relevant industry experience and strong written English. Work remotely on an async schedule for at least 10 hours each week, with no stated maximum.
Location and eligibility
Remote · Eligibility not specified
Engagement type
Not specified
Compensation
$115/hr
chemical engineering
mass balances
process design
EngineeringView job
Showing 20 active roles of 801. Dates reflect what the source published, not when AIWorkJobs last synced.
Sources: ten specialist AI work platforms. AIWorkJobs does not verify every claim made by a source. Report a listing that looks wrong through the contact page.