Role directory
Every kind of work that trains AI.
AI-training work is not one job. It spans rating model answers, writing expert reasoning, labeling data, lending a professional credential, and more, and the pay swings enormously by which one you do. Pick the work that fits your background to see what it pays and who hires for it, or browse by platform instead.
By kind of work
12 kinds of AI training work.
Each links to a full breakdown of what the work involves, what it pays across the market, and every platform that offers it, with its open roles right now.
Law, Medicine & Finance
1,486 open now55 platforms offer it
Credentialed professionals (physicians, attorneys, CPAs, financial analysts, consultants) evaluate model reasoning in their field. The highest-ceiling work in the industry: in 2026 the expert-data premium climbed steeply, with medical, legal, and finance specialists commanding $250–$450/hr and the rarest consulting and executive engagements reaching $500–$1,000/hr for a verifiable license and years of experience.
Live median $70/hr. Measured from 988 distinct advertised rates, across 38 platforms. 148 of the 1,486 open roles here state no hourly rate.
Multilingual & Localization
906 open now33 platforms offer it
Provide native-language data, translation, and dialect expertise, especially for the underserved and low-resource languages frontier labs are racing to cover as they close the multilingual reasoning gap. Native fluency in an in-demand language is a genuine premium skill, not just an entry ramp for non-US workers.
Live median $35/hr. Measured from 559 distinct advertised rates, across 22 platforms. 203 of the 906 open roles here state no hourly rate.
Transcription, Audio & Speech
610 open now31 platforms offer it
Transcribe, edit, and annotate recorded speech so AI models can hear and speak, or record your own voice for speech datasets. It spans accessible per-task transcription and voice recording and a fast-growing premium tier: native-language transcription, audio annotation, voice and audio evaluation, and AI-trainer sound engineering now command the top rates.
Live median $30/hr. Measured from 414 distinct advertised rates, across 21 platforms. 192 of the 610 open roles here state no hourly rate.
Coding & Software Evaluation
315 open now39 platforms offer it
Review, write, and grade code so models can program: bug-finding, code-generation review, agentic-coding evaluation, and worked engineering examples. One of the best-paid categories for people with real software experience.
Live median $60/hr. Measured from 210 distinct advertised rates, across 18 platforms. 33 of the 315 open roles here state no hourly rate.
Math & STEM
247 open now30 platforms offer it
Author and verify advanced math, physics, chemistry, and biology problems, including formal proofs. Rare expertise (formal-proof mathematics, physics PhDs) sits at the very top of the pay scale on several platforms.
Live median $40/hr. Measured from 212 distinct advertised rates, across 18 platforms. 18 of the 247 open roles here state no hourly rate.
Data Annotation & Labeling
236 open now42 platforms offer it
Annotate and label images, video, LiDAR point clouds, and text so models can perceive and categorize the world. Data annotation is the most accessible entry point into AI training, though the simplest labeling is increasingly automated by AI pre-labeling, so pay stays geography-priced with a real floor while budgets shift toward verifiable expert work.
Live median $30/hr. Measured from 114 distinct advertised rates, across 19 platforms. 71 of the 236 open roles here state no hourly rate.
AI Evaluation & Rating (RLHF)
145 open now63 platforms offer it
Rate, rank, and compare model responses so a model learns which answers are better. The backbone of modern alignment work, and an increasingly tiered one: routine rating is commoditizing while the premium moves to credentialed expert evaluation and reinforcement-learning reward signals.
Live median $40/hr. Measured from 101 distinct advertised rates, across 23 platforms. 33 of the 145 open roles here state no hourly rate.
Agentic Tasks & RL Environments
106 open now18 platforms offer it
Build and grade the interactive tasks and reinforcement-learning environments that agentic models train against: author multi-step tool-use scenarios, annotate agent trajectories and reasoning traces, and design verifiable-reward graders. This is the 2026 frontier of AI-training work and one of the best-paid, because it needs real domain judgment rather than volume.
Live median $60/hr. Measured from 64 distinct advertised rates, across 13 platforms. 21 of the 106 open roles here state no hourly rate.
Robotics & Physical-AI Data
77 open now6 platforms offer it
Generate and annotate the data that teaches robots and embodied AI to act in the physical world: label robot and egocentric video, annotate vision-language-action (VLA) demonstrations, collect teleoperation and manipulation data, and quality-check robot-perception datasets. This is the newest frontier of AI-training work, growing fast as humanoid and physical-AI labs race to close the real-world data gap that text and images cannot fill.
Red-Teaming & AI Safety
51 open now21 platforms offer it
Adversarially probe models to find unsafe, biased, or wrong behavior before release. Blends security, domain, and creative skills; pays a premium over generalist rating.
Live median $80/hr. Measured from 31 distinct advertised rates, across 11 platforms. 6 of the 51 open roles here state no hourly rate.
Writing & Content
37 open now17 platforms offer it
Write prompts, craft reference answers, and create or edit content that teaches models to write well. A strong fit for professional writers and educated generalists.
Live median $60/hr. Measured from 21 distinct advertised rates, across 7 platforms. 6 of the 37 open roles here state no hourly rate.
Research Studies & Surveys
22 open now9 platforms offer it
Take part in surveys, experiments, and human-preference studies that generate research and evaluation data. Two tiers: low-barrier consumer surveys are flexible supplemental income, while professional and expert panels (B2B interviews, paid expert consultations) pay materially more for verified expertise.
Live median $70/hr. Measured from 12 distinct advertised rates, across 7 platforms. 8 of the 22 open roles here state no hourly rate.
New to AI-training work? Read what an AI trainer does and how to become one, or start with the learning guides to understand what each kind of work involves and how to begin, then come back to compare the pay.
Keep exploring
Choose by platform, pay, or where to start.
The kind of work is one axis. Compare the platforms that hire for these roles, dig into the live pay numbers, or learn the ropes before you apply.
Browse by platform
All 80 platforms that pay you to train AI, with each one's pay, payout terms, legitimacy, and live open roles.
Explore platforms →Compare the pay
Live hourly rates across every kind of work and platform, plus an estimator for what your own background could earn.
Open the pay explorer →New to this?
Plain-language guides to every kind of AI-training work: what it is, who it suits, and exactly how to begin.
Start learning →