Rate, write and check AI answers for pay. Expert tracks pay more; most start with an unpaid assessment.
11 listed
AI training work
Remote freelance contributors write prompts, build grading rubrics, and rate or rank AI model responses on projects matched to their expertise. Work is paid per task-hour at a project rate shown before you start, settled weekly.
AI training work
Contractors evaluate chatbot responses, compare AI outputs, write and review code, or do domain-specific STEM/professional work to train AI models. Work is paid hourly by project track, and earnings are withdrawn to PayPal.
AI training work
Alignerr (Labelbox) contracts experts and generalists to rate AI outputs, annotate documents and evaluate code remotely. Work is paid hourly per project, with rates shown on each job listing and payouts made weekly.
AI training work
Mercor is a talent marketplace that matches professionals (doctors, lawyers, engineers, consultants) to hourly AI-training and evaluation contracts for AI labs. Contractors track hours and are paid weekly in USD.
AI training work
Handshake AI's fellowship (including the MOVE 'Model Validation Experts' track for master's/PhD holders) pays students and degree holders to evaluate and improve AI models in their discipline. Work is remote, part-time and paid hourly at a rate shown before you commit to a project.
AI training work
Invisible Technologies recruits freelance AI trainers through Meridial, its expert contractor marketplace, for projects writing, evaluating and correcting AI model output. Each project states its scope and rate in a Statement of Work, and contractors are typically paid hourly about twice a month.
AI training work
TELUS Digital (formerly TELUS International AI) hires part-time remote contractors for its AI Community to rate search results, ads and AI responses, review content and collect data. Pay is hourly or task-based by program, with estimated earnings listed on each posting.
AI training work
Appen recruits contributors through CrowdGen for projects such as search evaluation, data annotation, and expert AI evaluation. Pay is set per project and location, and contributors log hours or tasks and are paid monthly.
AI training work
Toloka runs Mindrift, a project-based platform where domain experts write prompts, evaluate AI outputs and improve responses; the legacy Toloka microtask app pays small amounts per task. Mindrift pays per completed task with the amount disclosed beforehand, processed bi-weekly.
AI training work
micro1 recruits contractors for AI-training, evaluation and engineering projects after an AI-led interview with its recruiter agent 'Zara'. Pay is hourly on approved tracked hours or per task, depending on the project.
AI training work
Turing hires software developers, data scientists and domain experts as contractors to generate training data, evaluate and improve LLMs for frontier AI labs, alongside traditional engineering roles. Contractors are paid an hourly or monthly contract rate in USD.
Pay ranges come from the platforms' own pages and reputable reviews, and vary by project, location and skill. Some links are referral links and are labelled where they appear; they never change what we list or the order we list it in.
