The work

You will evaluate research-level physics tasks and AI-generated answers in your area of specialization. Day-to-day work includes checking derivations, assumptions, dimensions, limiting cases, and physical interpretation; writing precise problem specifications and reference solutions; and identifying answers that sound credible but would not withstand expert review.

Evaluation and benchmark design

You will design challenging problems and evaluation sets that distinguish genuine physical reasoning from pattern matching. You may also calibrate judgments with researchers in adjacent fields, turning tacit standards about acceptable approximations, completeness, and correctness into explicit scoring criteria.

What the platform screens for

The screening is likely to verify your PhD, current or previous senior research appointment, publication record, and depth in a specific physics subfield. Expect follow-up questions about concrete research contributions, examples of flawed reasoning you have diagnosed, your use of large language models, and how you would grade technically plausible but incomplete solutions.

Logistics

The role is fully remote and limited to candidates based in the United States. The listing does not specify weekly hours, schedule, project duration, or whether the work is primarily asynchronous, so candidates should be ready to state their realistic availability and clarify scheduling expectations during screening.

Pay band: $90–140/hr