Research Physics Expert | $80-$135/hr Remote
Overview
We're assembling a team of research-level physicists to create golden reference solutions for the CritPt benchmark, a frontier physics reasoning dataset that tests the limits of large language models. You'll solve intricate physics problems from scratch, audit peers' work, or settle disputes between competing approaches — all done remotely and on your own schedule. This is a chance to contribute to AI evaluation with rigorous, human-verified physics, while earning $80–$135 per hour.
What You'll Do6
- 1Work through demanding physics research problems from start to finish, providing derivations, runnable code, and peer-reviewed citations
- 2Break each problem into smaller checkpoint tasks that force genuine physical insight, not just formula-crunching
- 3Develop Python answer templates with auto-grading logic for both symbolic and numerical responses
- 4Review other experts' solutions for mathematical correctness, methodological soundness, and missing edge cases, then give clear, iterative feedback
- 5When two solvers reach different answers, evaluate both paths and choose the one that becomes the trusted reference solution
- 6Document every step: chain-of-thought reasoning, accepted error tolerances, equivalent symbolic expressions, and the verification tests used
Requirements7
- 1For solver roles: an active PhD or postdoc in a relevant physics subfield — a senior doctoral student is the minimum bar
- 2For auditor roles: a completed PhD plus postdoc or junior faculty experience in the target area
- 3For adjudicator roles: a senior postdoc, junior professor, or industry research lead with a strong publication record
- 4Hands-on familiarity with at least two core methodologies of your specialty, proven through publications — broader coverage is a big plus
- 5Submit 3–5 representative papers (via arXiv IDs or DOIs), ideally from the last five years and in the subfield you're applying for
- 6Comfortable with LaTeX, Python, Jupyter, and SymPy for symbolic and numeric computation
- 7Strong written English at C1/C2 level, with native-level fluency a plus
Who Should Apply
You're the kind of physicist who genuinely enjoys solving hard, unstructured problems and insists on getting every detail right. Whether you're a graduate student, postdoc, or seasoned professor, you have a solid publication record in a relevant subdomain and can switch between analytical thinking and hands-on coding without missing a beat. You value precision, are comfortable explaining your reasoning in writing, and like the flexibility of asynchronous remote work.
Salary Insight
Pay is $80–$135 per hour, depending on the role (solver, auditor, or adjudicator) and the depth of your demonstrated expertise.
Location
Required Skills
Application Tip
In your cover letter, list the exact physics subdomains you're strongest in (e.g., quantum information or condensed matter) and include direct arXiv links to your 3–5 best papers — this will help match you to the right task pool quickly.
Similar open positions
Explore active roles that match your skills and interests.
Mercor
VerifiedPhysics Research Collaborator Remote
Mercor connects elite creative and technical talent with leading AI research labs. Headquartered in San Francisco our investors include Benchmark General Catalyst Peter Thiel Adam D'Angelo Larry Summers Jack Dorsey. This part-time position offers $80–$110/hour remote contract work.
Mercor
VerifiedComputational Chemistry & Electronic Structure Expert | $70-$100/hr Remote
Join a project that measures how well advanced AI systems tackle real scientific and engineering challenges. As a task designer, you'll craft original, graduate-level computational problems centered on real research workflows, then test them against leading AI models and fine-tune the difficulty until each one lands just right. We're currently seeking experts in computational chemistry and electronic structure—especially those with hands-on PySCF experience—to help build this large-scale benchmark. This is a remote, hourly role that blends deep scientific knowledge with puzzle-like problem design.
Mercor
VerifiedPhysicist Talent Network | $60-$80/hr Remote
This isn't a traditional job posting—it's an invitation to join Mercor's Physicist Expert Network, a community of physicists connected with leading AI labs and companies. As a member, you'll apply your physics expertise to train and evaluate AI models, design tasks based on real-world scenarios, and provide domain-specific feedback that pushes frontier AI research forward. You'll work remotely on a schedule that suits you, with typical commitments of 15-30 hours per week. What sets this apart: you apply once, go through a verification process, and get matched to contract opportunities as they arise—earning between $60-$80 per hour.
Mercor
VerifiedComputational Astrophysics & Cosmology Expert | $70-$100/hr Remote
This role is for a computational astrophysics and cosmology expert who will design graduate-level problems that test AI models' ability to use real scientific software. You'll craft challenging tasks, run them against state-of-the-art AI models, and iterate until the difficulty is calibrated. This is not data labeling — it's about creating problems that require deep domain expertise, strategic thinking, and hands-on use of tools like astropy. The work is fully remote, hourly, and flexible within a 15-20 hour weekly commitment.
Mercor
VerifiedBiophysics Researcher - Scientific Expert
Mercor connects elite creative and technical talent with leading AI research labs. Based in San Francisco, we partner with top investors including Benchmark and General Catalyst. This part-time position offers remote work up to 20 hours weekly.
Mercor
VerifiedData Science & Quantitative Analysis Expert | $60-$90/hr Remote
A premier AI research organization is assembling a team of quantitative experts to help design the next generation of evaluation benchmarks for frontier models. In this role, you'll create realistic, hands-on data analysis challenges that push models to their limits — cleaning messy datasets, comparing statistical methods, and verifying whether model outputs hold up to scrutiny. This is a fully remote, W-2 position with Cincinnatus LLC, working about 35 hours per week and collaborating closely with the lab's researchers. If you're passionate about the science behind AI and want your analytical expertise to directly shape how models are measured, this is a unique opportunity.