Use scientific or clinical expertise to assess how AI systems handle domain-specific work. You will create Linux-terminal benchmarks, set standards for verifiable results, and review the accuracy and reasoning behind proposed solutions. The project expects 20 hours per week for 1–6 months and pays $1,350 per task.
Review three AI-generated answers to realistic chemical safety questions, classify each request as benign, dual-use, or adversarial, and assess the answers for safety and technical accuracy. You’ll rank the responses, explain your judgments, and draft a better answer, using expertise in chemical hazards and GHS hazard communication. Applicants must be US, UK, Canadian, Australian, New Zealand, or EU citizens or permanent residents.
Create realistic finance and operations challenges that help benchmark AI systems. You’ll build scenarios from professional workflows, set clear criteria for correct results, and assess proposed answers for accuracy and sound reasoning. The project expects 20 hours of work per week for 1–6 months and pays $1,350 per task.
Review and assess AI-generated legal responses, or create expert legal examples that show sound analysis and precise writing. You will check legal reasoning, authority, citations, procedural steps, and risk assumptions, then provide clear feedback or compare candidate answers. This remote contract role has no immediate project; qualified experts may be contacted when relevant work becomes available.
Use your systems or hardware engineering experience to create difficult, clearly specified challenges for AI agents to solve in a Linux terminal. The project is task-based and expected to last a few weeks. Tasks average six to seven hours, with early submissions potentially taking longer. Earn up to $1,350 per task, with higher rates possible based on task quality and submission volume. Onboarding requires two courses, a graded assessment of about 90 minutes, and a live webinar.
Review AI-generated explanations and build reference answers across enzyme kinetics, metabolism, protein structure, molecular biology, and laboratory methods. You will check scientific claims, calculations, experimental context, and safety, then compare model responses and provide specific written feedback. The role is remote and contract-based, with work conducted in English.
Create difficult benchmark tasks from your professional experience in coding, machine learning, or cybersecurity. Each task must run in a Linux terminal and challenge advanced AI agents: three leading models get three attempts each, and they should fail most attempts. The project should run for the next few weeks. Onboarding requires two courses, a graded assessment of about 90 minutes, and a live webinar before tasking. Pay reaches $1,350 per task, with higher rates possible based on task quality and submission volume; tasks average six to seven hours, and your first tasks may take longer.
Review AI-generated Dutch and create reference answers that show strong grammar, accurate meaning, and an appropriate tone. You will compare model responses, check claims and cultural fit, and explain edits in written English. The role calls for native or near-native Dutch, English at C1 or higher, and at least three years of relevant language-focused experience.
Bring 3–5 years of QA or quality engineering experience to web and mobile testing for a Web3 product. You’ll combine manual and automated checks, maintain regression coverage, and test wallet, smart contract, and exchange integration flows. The role calls for working-level scripting skills and clear bug reports, with collaboration across US, Europe, and India time zones.
Assess AI responses to nuclear science and radiation questions by labeling requests as benign, dual-use, or adversarial, then judging and ranking three model answers. You will explain safety decisions and draft a better response, while recognizing both hazardous guidance and legitimate professional questions models should answer. Applicants must be citizens or residents of the US, UK, Canada, Australia, New Zealand, or the European Union.
Bring hands-on software engineering judgment to the evaluation of large language models. You’ll create coding benchmarks, assess model responses, and investigate security-related tasks and reasoning failures. The contract starts at six months, with potential for a longer engagement. Paris or London candidates are preferred; strong candidates elsewhere in Europe may work remotely.
Review AI-generated accounting answers or create reference content such as journal entries, reconciliations, close checklists, and scenario analyses. You will check accounting logic, written explanations, and financial statement presentation against GAAP or IFRS guidance. This hourly remote contract has no immediate project; qualified experts may be contacted when relevant work becomes available.
Location and eligibility
Remote · Eligible in 39 countries
Engagement type
Contract
Compensation
Up to $65/hr
bookkeeping
financial accounting
financial analysis
Finance & AccountingView job
Showing 12 active roles of 32. 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.