What the work involves
You write physics problems that a strong frontier model should plausibly fail, then supply the reference solution and the reasoning trace that shows exactly where the model goes wrong. In AMO and optical-materials work that usually means selection rules and transition strengths, laser cooling and trapping, coherent population dynamics, dielectric function and Kramers–Kronig relations, plasmonics, nonlinear susceptibilities, band-structure-derived absorption edges, and the spectroscopy that connects them to measurable data. A second stream of work is grading: you compare two model responses to the same prompt and justify a preference in written rubric terms — dimensional consistency, correct limiting behaviour, valid approximations, whether a plausible-looking numeric answer survives an order-of-magnitude check.
What the platform screens for
micro1's screening is AI-led and conversational. Expect it to ask what you actually did in your PhD or postdoc, then push one or two layers deeper into a specific mechanism — not to trap you, but because the fastest way to distinguish a genuine AMO physicist from a generalist is to follow up on a physical assumption. It also probes evaluation judgment: whether you can articulate why an answer is wrong rather than only that it is, and whether you can write a problem that is hard for the right reason instead of merely tedious. LaTeX fluency and clean, unambiguous problem statements matter more than publication count.
Logistics
- Fully remote and largely asynchronous; work is claimed from a queue rather than scheduled.
- Contributors typically commit 10–20 hours a week, though projects vary and volume can be uneven between them.
- Pay is hourly within an observed $100–170 band; placement reflects subfield fit and calibration performance, and no band is guaranteed for any individual.
- Output is reviewed. Sustained work usually depends on passing a calibration set and holding agreement with senior reviewers.