TrainBrain builds expert-authored training data and evaluations for AI labs. We're hiring people with real domain knowledge to write, solve, and grade the problems frontier models are measured against.

Domains we're hiring in

• Healthcare & Medicine — physicians, nurses, pharmacists, clinical researchers, and senior medical, nursing, or pharmacy students
• Law — attorneys, solicitors, paralegals, legal researchers, and law students with strong doctrinal grounding
• Finance & Accounting — accountants, analysts, controllers, auditors, and advanced finance or accounting students
• Software Engineering — working engineers and strong CS students across languages and problem domains

No AI background required. We train you on the tooling.

What you'd actually do

Every model that returns a broken function, a wrong reconciliation, a misread statute, or an unsafe clinical suggestion got there because no expert caught it during training. You'd be that expert.

• Write hard, realistic problems from your field — the kind that sit right at the edge of what current models can handle
• Solve them yourself, so there's a verifiable correct answer
• Grade model attempts and explain precisely where the reasoning breaks down
• Refine tasks until both the difficulty and the gradability are right

No clients, no month-end close, no tickets, no on-call. It's written work, on your own schedule.

Terms

• Pay: $45–$100/hr, set by domain, seniority, and assessment performance. Specialised and hard-to-source expertise sits at the top of the band. Paid on a regular cadence.
• Schedule: You choose your projects, hours, and volume. Scale up in quiet weeks, scale down when term or your day job gets busy.
• Growth: Strong contributors move into higher-rate specialist projects and review roles.