
AI Trainer (Mathematics & Statistics)
$60-120
/hr
Mathematics roles pay mathematicians for formal proof verification (using Lean 4), abstract algebra evaluation, calculus problem assessment, and discrete mathematics verification. You will help AI understand rigorous mathematical reasoning and logical proof. Requires advanced mathematics education (preferably PhD). Most roles pay $50-$118/hr.
Part of: STEM & Science
103 jobs
Page 1 of 4

Trending jobs, new guides and pay data from listings, in one weekly email.

Up to$75
/hr

Up to$70
/hr

Up to$25
/hr
One browser alert a day with new roles, Mathematics first. You don't need an email.
Or follow new roles in Telegram: @aitrainerwork
Mathematics evaluation spans everything from calculus and linear algebra to abstract algebra, topology, and formal proof verification. Tasks present AI-generated proofs or solutions and ask you to identify logical gaps, incorrect lemma applications, or algebraic errors. Lean 4 and Mathlib formalization tasks are an emerging and well-compensated subset, if you can write formal proofs in a proof assistant, premium tasks await. Combinatorics and number theory problems appear frequently, as do probability and statistics tasks. Some tasks ask you to write counter-examples to disprove AI conjectures. Work is self-paced and intellectually engaging. Graduate-level math background is typical for advanced tasks; solid undergraduate training is sufficient for foundational problem-solving roles.
Not always, but most math roles ask for a Master's or PhD. PhD holders qualify for the highest rates, and a few roles accept a Bachelor's with research or teaching experience.
Lean 4 is a theorem prover for formal mathematics. You don't need it to start, but the highest-paying formal verification roles use it, and free online courses cover the basics.
Abstract algebra, real analysis, topology and discrete mathematics come up most often. Formal verification roles that use Lean 4 pay the most, and applied math and numerical analysis are also in demand.
You grade an AI's proofs and solutions instead of teaching students. You spot errors, check rigor and judge whether the model's reasoning holds up.
Yes, the work is part-time and fits around a teaching or research schedule. It is contract work rather than a second employer, but check your institution's outside-work policy first.
There are 103 open listings, from 14 platforms: Meridial (20), AnyoneAI (19), Mercor (18), Alignerr (16) and Turing (10). The newest one was added 7 days ago. The count changes as platforms post and close roles.
Most pay $50 to $118 an hour, and typical pay is $80, based on 93 posted hourly rates. The highest rate currently posted here is $210 an hour. These are the rates the platforms posted. The rate you're offered depends on how you do in the platform's assessment.
Meridial (20), AnyoneAI (19), Mercor (18), Alignerr (16) and Turing (10). No single one dominates this category, so it is worth holding accounts on more than one. Most contracts allow you to work through several platforms at the same time.
How much do you enjoy aitrainer.work?
Thanks for telling us.
A short review on Trustpilot helps other AI trainers find us.