What the work actually involves
You'll be handed real inpatient records — often a mix of structured FHIR resources and free-text notes — alongside model output that summarizes, extracts, or reasons about them. Your job is verification: does the assessment the model produced follow from what's in the chart, or did it invent a diagnosis, drop a pertinent negative, or carry forward a resolved problem? A typical task might be checking whether a generated discharge summary's problem list matches the documented course, or flagging that an HPI attributes a symptom timeline the notes don't support. You'll also mark missing, internally inconsistent, or incorrect information in the source records themselves, and write short rationales that a project team and a model-training pipeline can both use.
What the screen is looking for
micro1's process is AI-led and leans on follow-up questioning. It is testing three things: that your inpatient experience is recent and real (you'll be asked about service lines, census, note types, EHR); that you can reason about documentation at the level of a chart reviewer, not just a treating clinician; and that you can hold a guideline steady when your clinical instinct disagrees with it. Candidates with CDI, utilization review, coding/QA, or peer-review backgrounds tend to screen well because they already think in terms of "supported by the record" rather than "clinically true."
Logistics
- Fully remote, contractor engagement, no patient contact and no licensure-dependent clinical decision-making.
- Largely asynchronous — tasks are queued and you work through them, with periodic calibration rounds or guideline updates.
- Hours are flexible and typically part-time; many reviewers fit this around clinical shifts, but sustained weekly availability matters more than large blocks.
- Observed pay band is $70–100/hr; actual rate depends on the project and your credentials, and is set by the platform rather than guaranteed here.