Mercor is recruiting high energy theory and nuclear theory PhDs to create, solve, review or audit research-level physics problems for CritPt, a public benchmark testing whether frontier models can do genuine physics research reasoning. Matching is narrow — you must be able to point to your own published papers on one of seven specific phenomena, from AdS/BCFT end-of-the-world branes to one-loop quasi-PDF matching to dark photon signals in interferometers.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

CritPt (described in [arXiv:2509.26574](https://arxiv.org/abs/2509.26574) — read it before applying) is a benchmark of research-level physics challenges. Depending on how your publication record sits relative to the seven panel areas, you may be authoring new problems in your subfield, producing reference solutions, reviewing another specialist's completed work, or auditing accepted items. The specific assignment is agreed with you after matching.

This is research-grade output, not volume annotation. Every deliverable has to be complete enough that another specialist in your subfield can follow the derivation and verify it independently, which means written reasoning is part of every assignment — not an optional appendix. Expect to work in LaTeX with Python, SymPy and Jupyter alongside; familiarity with an agentic coding extension in VS Code is useful but not required if you say so up front.

What the platform screens for

The phenomenon, not the neighbourhood. The single most common decline is an excellent physicist who commands the methods but has published on the adjacent problem. "High energy theory broadly" does not match; "one-loop matching kernels for quasi-PDFs in Coulomb gauge" does.
Verifiable authorship. Three to five representative papers with arXiv IDs or DOIs, ideally from the last five years, first author strongly preferred. Every listed paper is checked against the public record, and an area selection submitted without a paper as proof is not scored at all.
Loop experience for the quasi-PDF areas. Lattice QCD experience alone will not carry areas 2, 3 or 4 — the panel needs people who have done the one-loop computation.
Written English at B2 or above, including sustained technical argument.
A PhD in high energy theory, nuclear theory or a closely related field is a hard requirement. Postdocs, research scientists and junior faculty fit best; senior PhD students with a strong first-author record are welcome.

Logistics

Remote, fully asynchronous, no fixed hours. The commitment is roughly 10 hours per week sustained across an 8 to 10 week window, starting immediately. The application is a form (education, research experience, method self-attestation, up to five area selections each backed by your own paper), followed by publication verification, then either direct onboarding or a short live alignment call. A 30–45 minute assessment is added only where the panel needs it; most applicants will not see one. Observed rate is $80–110 per hour, set by depth of subdomain match — not guaranteed, and the top of the band goes to exact matches.

High Energy Physicist (PhD)

What the work involves

What the platform screens for

Logistics

Interview prep — 10 questions