What the work involves

You'll be handed AI outputs that look like the work product of a revenue integrity analyst: a flagged variance on a DRG-based inpatient claim, a reconciliation between expected and allowed amounts under a percent-of-billed-charges arrangement, a draft underpayment appeal letter to a Medicare Advantage plan. Your job is to decide whether each output is defensible against the contract language it claims to apply — and if it isn't, to say precisely where it broke. That usually means checking whether the model applied the correct fee schedule version, honored carve-outs and implant/high-cost-drug provisions, calculated outlier thresholds correctly, or hallucinated a payment methodology the contract doesn't contain.

Beyond pass/fail judgments, expect structured annotation work: rating outputs against rubrics, writing rationales a non-clinician reviewer can follow, and occasionally authoring gold-standard responses the model is trained toward. Some projects also ask contributors to write adversarial prompts — contract scenarios with stacked lesser-of provisions, retroactive rate changes, or timely-filing edge cases — designed to surface where the system is overconfident.

What the screen looks for

  • Verifiable operating history. Five-plus years in underpayment recovery, payment variance analysis, or managed care contracting, with specifics: contract types, payer mix, systems used, recovery dollars you can characterize honestly.
  • Methodology depth under follow-up. Expect the interview to push past definitions into how you'd model a per diem with a stop-loss, or reconcile a case-rate claim where the payer applied an outdated fee schedule.
  • Calibrated judgment. Reviewers who mark everything wrong are as unhelpful as reviewers who wave things through. The screen probes whether you can distinguish a material payment error from a stylistic difference in how an appeal is worded.
  • Written clarity. Feedback that trainers can act on, not just a verdict.

Logistics

Remote and asynchronous, with work drawn from a queue rather than scheduled shifts. Contributors typically commit 10–20 hours per week, though project volume fluctuates and some engagements ramp up or pause with little notice. Pay observed at $85/hr for this listing; Mercor sets rates per project and per contributor, so treat the band as a reference point rather than a guarantee. Engagements are 1099 contractor arrangements, and onboarding generally includes a calibration round before paid work begins.