For AI training companies
The experts your data work needs, found with the whole market's data.
AITraining.jobs reads every open AI training role on 29 platforms every day and sees who applies for them. We use it to find the experts you need, among people already looking for this work and professionals who are not, assess every profile against your brief, and introduce only people who said yes. You pay only for the experts you accept.
Applications counted by our own analytics, Sep 18 to Oct 17, 2026, automated traffic excluded.
Why it works
What we hold that an agency does not
The whole market, read daily
Every open role on 29 platforms: who hires which expertise, in which languages and countries, at what posted pay. So we know what your experts are offered elsewhere and what it takes to win them. The same data feeds our research reports.
Experts already come to us
6,396 applications through us in the last 30 days, from people choosing AI training work, with pages written for lawyers, physicians, psychologists, engineers and 37 languages.
We reach the ones who are not looking
For each brief our system reads the role, measures how many professionals in the United States fit it and can be reached, and assesses each one it finds, reading their full professional history for senior roles. Only people who match the brief hear from us, one by one, never in bulk.
We know who gets through
We judge applications against each role's requirements and follow people through 17 platforms' own referral dashboards, from sign-up to paid work. So we know which profiles pass which screens, and we bring you those.
How it works
From your brief to the experts you accept
Tell us who you need
The expertise, credentials, languages and countries, how many people and by when. A short chat or an email is enough.
We size it before anyone is contacted
How many such people exist and can be reached, what comparable roles pay on other platforms, and what usually stops this profile at screening.
We source and assess
People already applying for this work through us first, then professionals who are not looking, reached one by one. Every profile is assessed against your brief, and every person is asked about your project.
You review a shortlist
Experts who fit the brief and said yes, each with why they fit. Your own screen decides, and you pay only for the experts you accept.
Pricing
Per expert you accept
Set per brief by how rare the profile is. Nothing for the brief, the sizing or the shortlist, and no listing or posting fees.
Inbound
What people applying through us choose
Applications to a specific role in the last 30 days, by kind of work: 4,394 of the 6,396. The rest went to a platform's general sign-up.
- AI Evaluation & Rating (RLHF)18%
- Multilingual & Localization17%
- Transcription, Audio & Speech15%
- Law, Medicine & Finance14%
- Data Annotation & Labeling10%
- Coding & Software Evaluation9%
- Math & STEM5%
- Writing & Content4%
- Agentic Tasks & RL Environments3%
- Robotics & Physical-AI Data3%
- Red-Teaming & AI Safety2%
- Research Studies & Surveys<1%
Ground rules
What never changes
Professionals trust this site because of these rules, and that trust is why the experts we bring you answer us.
- Every legitimate platform is listed, whether it pays us or not.
- Pay figures are never for sale. Pay tables and pay sorts rank by pay alone.
- We never sell candidate data. A person's details reach you only after they said yes to your project.
- We write to professionals one by one about one project that fits them, never in bulk.
How partner status affects what job seekers see is published in our methodology.
Questions
What companies ask us
Are you a recruiting agency?
No. We run a live map of the AI training labor market: every open role on 29 platforms hiring, the people applying for them through us, and how they fare at each platform's screen. Sourcing is what that data makes possible: we know before we start how many people fit a brief, where they are and what they are offered elsewhere.
Where do the experts come from?
Two places. People already looking for AI training work, who come to us from search and AI assistants and apply through us. And professionals who are not looking: for each brief our system finds the people whose own profile fits it and writes to them one by one. That outreach is to professionals in the United States.
How do you assess them?
Against your brief: their profession, the credentials and licences it requires, their specialty and seniority, and, for senior roles, their full professional history. What our judgments of past applicants in the same field show (the gaps that most often fail a screen) is part of the assessment. Your own interviews, tests and ID checks still decide.
What does it cost?
A fee per expert you accept, set per brief by how rare the profile is. Nothing for the brief, the sizing or the shortlist, and no listing or posting fees.
Which experts do you source best?
Experts whose time is scarce: lawyers, physicians and clinicians, psychologists, engineers, data scientists, mathematicians and finance professionals, and native speakers of languages that are hard to find. These are the roles where the data and the assessment matter most.
We need training data, not experts. Can you help?
Yes, through AIData.Broker, which the team behind AITraining.jobs also runs: tell it what data you need and it sources the data to your spec, for labs, data teams and companies training their own models.
Can we use your data without a hire?
Yes. Our research reports, pay figures and role data are public, and the records behind them, custom cuts and regular feeds are available on request.
Tell us who you need
A short chat, or an email. We reply with how many such experts there are, what they are offered elsewhere, and how we would reach them.
Talk to Yuma Heymans, Founder & CEO of AITraining.jobs.
Founder and CEO of AITraining.jobs and co-founder of the AI recruitment platform HeroHunt.ai, based in San Francisco, with years spent mapping how AI-training and data work is sourced, staffed, and paid.