US Contract Attorney — Transactional
$100–$120/hrRemote (United States)Contract
Apply now →You'll finish your application on Mercor, our hiring partner: sign up and take their short AI interview.
About the Role Mercor is partnering with a leading AI lab to train frontier models on high-quality legal reasoning data. This is an evergreen talent pool for experienced US Transactional attorneys — by applying, you join a standing bench and are considered for ongoing and upcoming legal-AI projects as they launch. We review applications on a rolling basis and reach out when work that fits your background comes up.
Typical work includes recurring side-by-side evaluations and legal dataset generation: you'll compare AI-generated legal work, judge which response is stronger and why, and create realistic transactional tasks and reference answers.
What You'll Do
Compare pairs of responses to transactional tasks (contract drafting, redlines, issues lists, due diligence summaries and deal analyses) and decide which is stronger, with clear reasoning
Write concise justifications grounded in the source documents and applicable US law
Create realistic legal prompts, fact patterns and model answers drawn from your transactional experience
Flag errors, gaps and risky advice in AI-generated legal content
Attend onboarding and calibration sessions
You're a Good Fit If You
Are admitted to practice in at least one US state (active bar membership)
Have 3+ years of US transactional practice, such as M&A, finance, capital markets, commercial contracts or corporate matters
Hold a J.D. from an ABA-accredited law school
Write clearly and pay close attention to detail
Bonus Qualifications
2+ years of experience at an AmLaw 100 firm (strongly preferred)
Prior work on legal AI projects with Mercor
Prior data annotation, AI evaluation or legal-benchmark experience
Role Highlights
Evergreen talent pool — apply once and be considered for ongoing projects as they launch
Fully remote, flexible hours; recurring tasks as new work comes up
Applications reviewed on a rolling basis