What the work involves

You will work against a queue of biology prompts — molecular and cell biology, genetics, biochemistry, microbiology, physiology, or evolutionary biology, depending on where your training sits. Typical tasks include comparing two model responses and justifying which is stronger, correcting a response that is fluent but wrong about a mechanism, and authoring new questions with reference answers that a strong model should plausibly fail. Written rationale matters as much as the score: reviewers downstream need to understand why a claim about, say, a signaling cascade or a knockout phenotype is incorrect, with a citation or a mechanistic argument rather than an assertion.

Because the category is labeled engineering, some batches lean quantitative — interpreting figures, evaluating experimental design, checking statistical reasoning in a methods description, or assessing bioinformatics and sequence-analysis steps. Expect calibration rounds early on, where your scores are compared against a rubric and other experts, followed by feedback and possible re-calibration.

What the platform screens for

  • A verifiable PhD (or near-completion) in a biology discipline, with publications or thesis work you can describe in detail.
  • Depth under follow-up: micro1's AI interviewer probes past your first answer, so a confident but shallow explanation of a mechanism tends to surface quickly.
  • Clear technical writing in English — concise, unhedged, and specific.
  • Judgment about uncertainty: knowing when a question is genuinely contested in the literature and saying so.

Logistics

Fully remote, asynchronous, contractor basis. Volume fluctuates with client demand; many experts treat this as 5–20 hours per week alongside academic or industry work, with some projects offering more during ramp-up. Pay is stated as observed for this listing, not guaranteed — actual rates vary by project, task type, and calibration performance.