What the work involves

Once you're in the pool, projects vary but the core task is consistent: put pressure on a model's clinical reasoning and document precisely where it breaks. That means writing prompts hard enough to separate a model that pattern-matches from one that reasons — a patient with two plausible diagnoses distinguished only by a single lab trend, a drug interaction that matters only at renal impairment, a nursing scenario where the safe action contradicts the obvious one. It also means grading: reading a model's answer against a rubric, marking accuracy, and writing a rationale that a non-clinician reviewer can audit.

The errors you're hunting are rarely blatant. A model quotes a potassium of 5.9 and then treats it as normal. A differential omits the one diagnosis that would kill the patient this shift. A dosing recommendation is correct for adults and silently wrong for a 70 kg adolescent. Guidance is cited confidently but is two revisions out of date. Your value is catching those at speed and explaining why they matter clinically, not just flagging that something looks off.

What the assessment screens for

AfterQuery's qualifier is timed, unpaid, and largely task-based rather than conversational. Expect to be shown model outputs and asked to identify errors, justify severity, and sometimes author a prompt of your own. Screeners look for: real clinical grounding under follow-up (vague answers get probed), calibration — you flag what is actually wrong and don't manufacture problems to look thorough — and written rationales that are specific, short, and defensible. Credential verification (licence, registration number, specialty) is typically part of onboarding into the paid pool rather than the assessment itself.

Logistics

  • Fully remote and asynchronous; work is claimed from a queue rather than scheduled.
  • Volume is uneven — busy stretches followed by quiet ones. Most contributors treat it as a supplement to clinical practice, commonly 5–15 hrs/week.
  • Rates vary by project difficulty and specialty scarcity; the $50–200/hr band reflects observed rates across matched projects, not a guarantee.
  • The assessment is unpaid and there is no obligation to accept projects once you pass.