High Energy Physicist (PhD)

Contribute to a public benchmark of research-level physics challenges designed to evaluate whether frontier AI systems can perform genuine research reasoning. This is research-grade work requiring complete, clearly written solutions or evaluations that another specialist can independently follow and verify. Review the benchmark paper, arXiv:2509.26574, before applying to assess your interest in the work. Role Overview

Assignments are matched to your published subfield expertise and may involve creating research problems, solving problems, reviewing completed work, or auditing it. The specific area and assignment are agreed after matching.

• AdS/BCFT, end-of-the-world branes, and black hole interiors: AdS/BCFT correspondence, Karch-Randall branes, geodesic approximations for heavy bulk fields, thermal one-point functions, BTZ geometry, and behind-horizon interiors.
• LaMET and quasi-PDFs, matching kernels and DGLAP evolution: pion PDFs, perturbative matching kernels, modified minimal subtraction renormalization, plus distributions, DGLAP evolution, renormalization-group resummation, and one-loop running coupling.
• LaMET and quasi-PDFs, one-loop perturbative QCD and renormalization: one-loop Feynman-diagram calculations, dimensional regularization, MS-bar renormalization, superficial degrees of divergence, and soft and collinear infrared divergences.
• LaMET and quasi-PDFs, Coulomb-gauge matching and lattice QCD parton structure: Coulomb gauge fixing, matching kernels, one-loop QCD corrections, dimensional regularization, MS-bar renormalization, and lattice-QCD parton-structure calculations.
• Ultralight vector dark matter in laser interferometers: dark photons coupled to the baryon-minus-lepton current, interferometric gravitational-wave detection, composition-dependent differential test-mass forces, strain-noise spectral density, detector sensitivity curves, coherent integration, signal-to-noise scaling, local dark-matter normalization, and stochastic-field coherence time.
• Matrix models and invariant theory: adjoint matrix models, Grassmann-valued invariants, gauge-invariant operator counting, trace relations, Cayley-Hamilton identities, Molien-Weyl formulas, plethystic Hilbert series, and primary and secondary invariants.
• Superconformal and Witten indices, finite-N trace relations, and plethystic counting: adjoint fermion matrix models, Grassmann-valued invariants, Molien-Weyl formulas, gauge-invariant counting, and finite-rank corrections to large-N counting.

• PhD in high energy theory, nuclear theory, or a closely related field. Postdoctoral researchers, research scientists, and junior faculty are especially well suited; senior PhD students with a strong first-author publication record may apply.
• Published work on the exact phenomenon for the selected research area, rather than only an adjacent topic. For quasi-PDF work, loop-calculation experience is required; lattice experience alone is insufficient.
• Publications demonstrating at least one relevant method family: holographic methods, perturbative QCD, effective field theory, lattice QCD, group-theoretic counting methods, or astroparticle and detector methods.
• A verifiable record of three to five representative papers, preferably from the past five years, with arXiv IDs or DOIs. First authorship is strongly preferred, and listed papers will be checked against the public record.
• Working proficiency with LaTeX, Python, SymPy, and Jupyter. Familiarity with an agentic coding extension in VS Code is helpful; candidates may disclose skill gaps honestly.
• English proficiency at B2 level or above, including the ability to communicate written reasoning clearly.

• Remote, asynchronous hourly engagement with no fixed working hours.
• Commit approximately 10 hours per week across an 8 to 10 week period, beginning immediately.

$80 to $110 per hour, determined by the depth of fit with the relevant subdomain.

• Complete the application form with education, research experience, a self-attestation of relevant methods, and up to five best-fit research areas.
• For every selected area, provide an arXiv ID or DOI for your own supporting paper, your author position, and the methods demonstrated. Selections without supporting evidence are not scored.
• Your papers and authorship will be verified against public records.
• Qualified candidates are either matched directly to an assignment or invited to a short live alignment call to confirm the research area and assignment. A 30 to 45 minute assessment may be requested when needed, though most applicants will not complete one.