What the work involves

You write the cases and then judge the machine on them. A typical task starts with a scenario you author from the bench outward: a specimen arrives, something about it is off or ambiguous, and the model has to reason through pre-analytical acceptability, instrument flags, delta checks, reflex testing, and what can defensibly be reported. You then write the reference answer — not just the correct result call, but the analytical reasoning that gets there, including the wrong turns a plausible-sounding responder would take.

The grading half is where the real signal lives. You read model output and mark it for analytical accuracy and patient-safety consequence, and the most valuable annotations are the ones that catch an interpretation that reads fluently, cites the right reference interval, and would still release a bad result — a cold agglutinin misread as an MCHC problem, an interference not considered, a unit that gets crossmatched on incomplete antibody workup. Written rationale matters as much as the verdict; downstream trainers use your reasoning, not your score.

What the screen looks for

  • Verifiable credential and department depth. MLS(ASCP) or equivalent, plus at least one department you can be pushed hard on — heme, chemistry, micro, or blood bank. Expect follow-ups that go past textbook level into what you actually do at 2 a.m. with a flagged instrument.
  • Pre-analytical and QA instinct. Whether you think about specimen integrity, calibration, and validation unprompted, or only about the number on the screen.
  • Calibrated grading. Screeners test whether you can distinguish a stylistic weakness from a reportable error, and whether you say "insufficient information" when a scenario genuinely lacks it.
  • Realistic capacity. 10 hours per week is a floor, not a target to negotiate downward.

Logistics

Fully remote, fully async — no shifts, no on-call, no third shift. You set hours each week, invoicing is hourly and paid weekly via Stripe, and project work is ongoing rather than a fixed engagement. Pay is stated as observed on the platform at $65–95/hr, with placement in that band typically tracking specialist certification and department depth rather than tenure alone.