Licensed Attorney — Legal AI Dataset Reviewer (Part-Time, Remote, Ongoing)

We build specialized, high-accuracy datasets used to train and evaluate legal AI systems. We're looking for a licensed U.S. attorney to be the quality anchor on a legal citation-verification project, setting the standard, reviewing annotations for legal accuracy, and resolving the tricky calls. This is detail-intensive work where your legal judgment is the deliverable. You won't be drafting briefs, you'll be judging whether legal propositions are genuinely supported by their citations, and defining the bar our annotators follow. What you'll do: Review and verify legal citation annotations for accuracy and consistency against written guidelines Refine those guidelines and set "gold-standard" answers Adjudicate disagreements and flag ambiguous cases Spot-check annotators' work and give correction feedback You're a strong fit if you have: A JD and active U.S. bar admission 3+ years in litigation, contracts, or legal research Meticulous attention to detail and the ability to explain why a citation is right or wrong Comfort with spreadsheets, digital tools, and fast feedback loops A reliable, self-directed work style Nice to have: familiarity with Bluebook citation, legal research databases (CourtListener, Westlaw, Lexis), or an interest in AI / legal tech. How we hire: shortlisted candidates complete a short paid test task (~30 minutes) — labeling a small set of real citations — before we finalize. It's the best predictor of fit, and yes, we pay for your time on it. Details: Hourly, remote, ~10 hrs/week to start, scaling with project volume. Strong potential for ongoing work.