AI labs pay scientists to train their models
Frontier AI companies hire PhDs and domain experts — biology, chemistry, physics, math, medicine — to write, grade and stress-test model answers in their field. It is real, flexible, remote contract work, and between positions it pays better than adjuncting. It is not research, and it is not a career; it is a bridge.
What the work actually is
Writing expert answers to hard questions, ranking model outputs, red-teaming claims in your specialty, building rubrics. Paid hourly, project-based, fully remote, on your schedule. Expect an application, a screening task, and inconsistent project availability — work comes in waves.
Real rates
Generalist work starts near $15–30/hr; PhD-level domain work commonly runs $30–75/hr, and scarce specialties (medicine, advanced math, niche lab expertise) reach $100+/hr. Rates are set per project — treat any number as the platform's ceiling claim until you see your own offer.
The honest downsides
Contract work with no benefits and no continuity guarantee. Projects pause without notice. NDAs are standard. It produces no publications and no first-author anything — budget it as income, not as career progress.
Where to sign up
Referral linksDisclosure: both links below are referral links — if you sign up and work, this site earns a referral fee at no cost to you. That relationship does not change the assessment above.
Looking for an actual position instead? The board lists live science roles — funded employers first.