What the work involves
You are handed model responses in your area of doctoral training and asked whether they would survive scrutiny from someone who actually knows the field. That means more than checking facts: it means identifying reasoning that reaches a correct conclusion through an invalid path, citations that look plausible but do not exist, arguments that skip the step a referee would demand, and answers that are technically accurate yet shallow enough to mislead. Alongside evaluation, you will write expert-level prompts designed to stress the model, author reference answers that define what a good response looks like, and produce critiques detailed enough that a lab engineer can act on them.
The field range is deliberately wide. Technical PhDs (mathematics, physics, chemistry, biology, engineering, computer science, economics) sit alongside English, literature, and journalism doctorates. In humanities work the failure modes differ — misattributed quotations, invented scholarly consensus, confident readings that ignore the text — and the evaluation rubric reflects that.
What the screening looks for
micro1's screen is AI-led and follow-up heavy. It will ask you to name your dissertation area and then push one or two layers deeper than a generalist could, looking for the specificity that only comes from having done the research: methods you actually ran, debates you actually have a position on, the paper you would cite. It also tests evaluation judgment — whether you can separate "wrong" from "unsupported" from "correct but hollow," and whether your written critique is concrete enough to be useful. Written English is assessed directly from your answers, not claimed.
Logistics
- Remote, contractor, asynchronous within project deadlines
- Project-based volume; expect it to fluctuate rather than run as steady full-time work
- Rates are observed in the $245–280/hr band and vary by field scarcity and task type, not guaranteed
- Preference for candidates based in an English-speaking country