What the work actually involves

You pick a subject you genuinely know — your thesis area, your professional specialty, the corner of a field where you can tell a correct answer from a plausible one — and build questions that a strong model should fail. The bar is not trivia. A usable item typically requires two or more retrieval steps plus a synthesis or inference the model has to perform itself, and it must have one defensible answer you can support with triangulated sources. You then run the question against the model, read the failure mode carefully, and adjust: sometimes the question was ambiguous rather than hard, sometimes the model got there by luck, sometimes it needs another reasoning hop. Every submission travels with citations and a written explanation of why the answer is what it is.

Expect reviewer feedback loops. Items get returned for ambiguity, for a citation that doesn't actually support the claim, for a question whose answer shifted since the source was published, or for difficulty that came from obscurity rather than reasoning. Rework is part of the job, not a sign something went wrong.

What the screen looks for

micro1's screening is AI-led and pushes on depth. It wants to hear a real domain — named coursework, a research topic, a professional practice area — and it will follow up inside whatever you name. It also probes research hygiene: how you verify a claim, what you do when two credible sources disagree, how you decide a question has exactly one answer. Prior AI-training experience is explicitly not required; demonstrated research and writing ability is. Recent graduates and advanced-degree holders are both in scope.

Logistics

  • Fully remote, asynchronous, contractor engagement.
  • Compensation is output-based — paid per task that meets spec. Observed bands for this listing run $20–60/hr equivalent; actual earnings depend on how fast you produce accepted work, and are not guaranteed.
  • A weekly minimum number of submissions applies.
  • Roles are typically filled within about 48 hours, with first tasks expected 24–48 hours after onboarding.
  • Written English fluency is required; it need not be your first language.