$80–110/hr
The role in one line
A research physicist with PhD-level expertise who creates and evaluates frontier physics problems in gravitation theory and high-energy astrophysics.
Written by Training Turk from the public listing; it may be incomplete or out of date. Read the full posting on Mercor.
What you would do
- Write original research-level physics problems from your own work, including theoretical foundations, solution pathways, and evaluation criteria
- Solve complex problems in torsion-based gravity, topological terms, cosmological perturbations, or relativistic astrophysics with complete documentation of your methodology
- Assess whether AI-proposed solutions exhibit sound physics reasoning and check for common conceptual errors in the phenomena you study
Who they are looking for
- PhD in gravitation physics, cosmology, astroparticle physics, or closely related advanced physics subfield
- Published research on specific phenomena: torsion, Nieh-Yan topological terms, axion inflation, or related topics
- Verifiable publication history of 3-5 representative papers from the last 5 years, with arXiv IDs or DOIs; every paper will be verified against the public record
- Advanced proficiency with LaTeX, Python, SymPy, and Jupyter notebooks; working knowledge of tetrad and spin-connection sign conventions and sign-convention discipline
What the interview is likely to probe
1.Specific subfield publication record
Researchers working in adjacent areas lack the depth to distinguish valid approaches from plausible-sounding errors; your papers directly prove your expertise in the exact phenomenon being tested.
Expect something like: “You apply to work on Nieh-Yan axion inflation. Describe your understanding of how topological terms couple to inflation dynamics and walk through a novel problem you could create.”
2.Problem authorship and completeness
A well-constructed research problem requires not just a question but a complete solution that an AI model's attempt can be graded against; incomplete reasoning invalidates the benchmark.
Expect something like: “Describe an original research problem from your recent work. What makes it challenging? How would you structure it for someone unfamiliar with your subfield?”
3.Solution verification and rigor checking
Your ability to identify where a model's reasoning is mathematically sound but physically unrealistic or incomplete separates valid evaluations from false positives.
Expect something like: “An AI proposes a synthesis route using thermodynamically allowed reagents and standard mechanisms. Walk through whether the approach is sound for your physics problem.”
4.Numerical method command and reproducibility
If you implement coupled field equations differently than a colleague, different sign conventions or integration schemes can yield disagreement; your explanations must be unambiguous enough to resolve such ambiguities.
Expect something like: “You're integrating background field equations in Palatini formulation with pre-inflationary dynamics. Explain your approach and key assumptions you'd want candidates to verify.”
5.Frontier AI reasoning beyond textbooks
The goal is to test whether AI can carry out research reasoning, not memorize known results; your problems must target gaps that textbooks and standard references don't cover.
Expect something like: “Describe a research question in your subfield that hasn't been thoroughly addressed in the literature you've read. How would you structure a problem around it?”
Exercise you may get
Write a complete research-level physics problem from your subfield with theoretical setup, solution pathway, and evaluation criteria for assessing candidate understanding.
How to prepare
- Identify your three strongest first-author papers and prepare a detailed technical summary of each, highlighting which specific methods and phenomena each demonstrates
- Gather published research on your specific topics to reference when creating interview questions
- Prepare a research problem from your own work that you believe tests AI reasoning about frontier physics rather than textbook knowledge
Facts
- Pay
- $80–110/hr
- Commitment
- hourly
- Hours
- 10 per week
- Work arrangement
- remote · Remote
- Domain
- Life, Physical, and Social Science
- Posted
- 9/25/2026
- Open slots
- 3