What the work actually involves

Most of your time is spent reading — AI-generated nursing notes, assessment summaries, handoff text, patient-facing responses — and judging whether they'd survive contact with a real shift. You mark what's wrong, but more importantly you say why: a missing neuro check in a post-op assessment, a pain reassessment that never closes the loop, a fall-risk score that doesn't match the narrative, a med entry that reads plausibly but isn't how anyone charts it. Other projects run the other direction: you write or structure gold-standard documentation from case material so a model has something correct to learn from.

A second stream of work is less scripted. You'll get ambiguous clinical questions — is this phrasing a safety issue or a style preference? does this output violate documentation standards or just deviate from one hospital's habit? — and be asked to investigate and return a recommendation. Strong annotators also push back on the guidelines themselves when a taxonomy doesn't fit what bedside charting actually looks like. That feedback loop is a real part of the job, not a courtesy.

What the screen is looking for

  • Currency over tenure. The listing says it outright: recent inpatient bedside documentation experience within the last five years matters more than total years in nursing. Expect questions that date your last bedside shift and probe which EHR you charted in.
  • Breadth, not subspecialty depth. You'll review across med-surg, step-down, ICU, telemetry, post-op. A candidate who can only speak to one narrow population is a harder fit than a floating RN with wide exposure.
  • Consistency under a rubric. Can you apply someone else's annotation guideline the same way on item 400 as on item 4, including when you personally disagree with it?
  • Licensure geography. Active U.S. RN license, outside California. This is a firm gate on Mercor's side, not a preference.

Logistics

Fully remote, asynchronous, project-based. The floor is 10 hours per week, reliably — sporadic bursts don't work because annotation quality is tracked against calibration sets over time. Hours are yours to schedule, which makes this workable alongside a clinical FTE or PRN line. Onboarding typically includes a paid or unpaid calibration task; ongoing work depends on agreement rates with reviewers. Rates of $55–65/hr reflect what contributors have reported for this role and are not guaranteed for any individual project.