General Science Expert (Train AI Models Part Time!)

About the Role

Mercor is partnering with a leading AI lab to strengthen expert-level scientific reasoning in frontier models. We are hiring broad-based science generalists to author and review challenging cross-disciplinary STEM problems and to evaluate AI-generated solutions and multi-step reasoning trajectories for correctness and rigor.

Responsibilities

• Design expert-level problems spanning multiple scientific disciplines (physics, chemistry, biology, mathematics), including olympiad-style and agentic scientific-discovery tasks
• Review problems authored by peers for clarity, genuine difficulty, and ground-truth correctness
• Evaluate and compare AI model outputs across domains, delivering Accept / Revise / Reject verdicts with detailed written rationale

Qualifications

• Demonstrated breadth and rigor across multiple STEM fields; science-olympiad medalists (IPhO/IChO/IBO/IMO), competition winners, or polymath researchers strongly preferred
• Ability to judge correctness outside a single specialty and to reason through multi-step scientific problems
• Strong scientific writing and meticulous attention to detail