What the work involves

You build "worlds" — self-contained regulatory scenarios that an AI model is asked to work through, paired with a reference answer you draft yourself and a rubric that explains what separates competent regulatory judgment from plausible-sounding filler. A typical task might be a compliance program redesign after an enforcement referral, a response to an agency information request, a comment letter on a proposed rule, or a risk memo on a novel product feature under an ambiguous regulation. Subject matter spans financial services, healthcare and FDA, antitrust and competition, privacy and cybersecurity, and environmental and energy regulation, and scenarios often reference the tooling real programs run on: regulatory tracking platforms, compliance management systems, and agency filing portals.

The rubric is the deliverable that matters most. Mercor is trying to capture how senior regulatory counsel actually reason — where the deference question lives, when a voluntary disclosure is the wrong call, why a technically correct citation still produces bad advice — rather than whether a model can recite the elements of a statute. Reviewers push back on rubrics whose criteria could be satisfied by generic bar-exam recall.

What the platform screens for

  • Verifiable ownership. Five-plus years as regulatory counsel or compliance counsel at a major firm, a regulated company, or an agency — and specific matters you personally drove, not matters you staffed.
  • Depth under follow-up. The AI interview asks a second and third question about the same regime; general familiarity with, say, the APA rulemaking sequence tends to surface quickly.
  • Credentials. JD with active bar admission or an international equivalent is strongly preferred; prior rubric, training-material, or compliance-documentation authorship helps.
  • Track fit. Say plainly whether you are US, International, or both. Overclaiming a second track is a common disqualifier.

Logistics

Fully remote, asynchronous, contractor engagement with no fixed hours. Most contributors take on a defined batch of worlds with a delivery window rather than a weekly quota; sustained contributors often run 10–20 hours a week. Pay in the $90–100/hr band reflects rates observed on recent Mercor regulatory listings and is not guaranteed — actual offers vary by track, specialty scarcity, and screening outcome. Expect a revision cycle on your first batch as you calibrate to house rubric conventions.