What the work involves
You build the material AI models are tested against. A typical task starts with a scenario drawn from work you've actually done — a DPA negotiation with a cross-border processor, an FCPA red flag surfacing mid-diligence, a SOX controls gap that has to reach the audit committee, an MSA indemnity stack that doesn't survive a change-of-control. You draft the scenario with enough factual texture that shortcuts fail, produce a reference answer at the standard you'd send to a GC, then write a criterion-referenced rubric that tells a grader exactly which analytical moves earn credit and which confident-sounding answers should lose points.
The rubric is the deliverable that matters most. Reviewers look for criteria tied to specific legal reasoning — identifying the controlling regulation, spotting the risk the client didn't ask about, sequencing remediation correctly — rather than vague quality bands. Expect work across in-house counsel, regulatory, transactional, and litigation tracks, plus ERM and COSO-aligned risk scenarios and board-level reporting.
What the screen looks for
- Verifiable enterprise scale. Two-plus years in-house at a Fortune 500 or large public company, or partner/senior associate at an AmLaw 100 firm serving F500 clients. Matter size, counterparty type, and deal or regulatory context get probed.
- Depth in a named practice area. M&A, securities, regulatory, employment, IP, privacy, or commercial contracting — the screen follows up on one and pushes for specifics.
- Active bar admission and a JD. Non-negotiable.
- Evaluation instinct. Whether you can articulate why one legal answer is better than another in terms someone else could apply consistently.
Logistics
Fully remote and asynchronous, with a 20 hour per week minimum commitment. Onboarding runs through the Mercor Rubric Academy, a paid calibration program you complete before live work. Observed rates run $110–150/hr, set by Mercor based on demonstrated expertise and not guaranteed for any individual contributor. Strong contributors are moved into reviewer, lead, and domain SME roles at elevated rates. Familiarity with CLM platforms (Ironclad, Agiloft, Icertis) and enterprise GRC tooling helps; prior rubric, bar exam, or CLE content authorship is a meaningful plus.