What the work involves

You will be assigned to one of five research areas and given one role within it: writing new research-level problems, solving problems written by others, reviewing completed solutions, or auditing finished items. The assignment is agreed after matching, and it follows from where your publications sit. In every case the output is written physics — a derivation another specialist in your subfield can follow and verify without talking to you. That means stated assumptions, intermediate steps that survive checking, and a clear account of why a particular method (duality, replica continuation, a BPZ null-vector equation, a saddle-point expansion) is the right one for the problem. This is research-grade output measured per item, not a volume annotation task.

The five areas are: disordered spin models and the Nishimori line, QEC thresholds and mixed-state topological order; entanglement entropy via the replica trick with Goldstone modes and tower of states; range expansions with long-range dispersal and heavy-tailed jumps; lattice path enumeration and entanglement domain-wall statistical mechanics; and two-dimensional CFT covering the Ising minimal model, BPZ equations, conformal blocks and Coulomb gas. The benchmark paper is arXiv:2509.26574 and the listing recommends reading it before applying — it is the fastest way to decide whether the work suits you.

What the platform screens for

  • Publication on the phenomenon, not the neighbourhood. The listing names this as the single most common reason strong physicists are declined. Command of the transfer matrix does not substitute for having published on entanglement domain-wall statistics.
  • Verifiable authorship. Three to five representative papers with arXiv IDs or DOIs, ideally within five years, first author strongly preferred. Area selections submitted without a paper as proof are not scored at all.
  • Written reasoning in English at B2 or above. A large share of the value is how legibly you set out an argument.
  • Tooling honesty. LaTeX, Python, SymPy and Jupyter are expected working tools; some exposure to an agentic coding extension in VS Code helps. Declared gaps are acceptable, overstated fluency is not.

After the form, papers are verified against the public record. You are then either onboarded directly or invited to a short live alignment call to fix the area and assignment. A 30–45 minute assessment may be added where the reviewers need it, but most applicants will not see one.

Logistics

Fully remote and asynchronous with no fixed hours. The commitment is roughly 10 hours per week sustained over an 8 to 10 week window, starting immediately. The observed band is $80–110 per hour, set by how closely your subdomain matches the area — the top of the band goes to people whose own papers sit directly on the listed phenomenon.