What the work actually involves

You will spend most of your time doing two things: producing gold-standard redlines on contracts you would recognise from any SaaS legal team — MSAs, NDAs, DPAs, order forms, reseller agreements — and grading what a model produced against them. A typical task hands you a counterparty draft, a fact pattern about the company's posture (which side you're on, risk tolerance, deal size), and a model's proposed markup. You assess whether the model spotted the uncapped indemnity, whether its limitation-of-liability rewrite is actually enforceable, whether it flagged a sub-processor clause that conflicts with the DPA, and whether its explanatory comment to the business would survive contact with a real client.

A second stream is framework work: writing the grading rubrics and issue-spotting checklists other reviewers use, so that scoring is consistent across dozens of annotators rather than a matter of taste. This is where in-house experience earns its premium — you know which clauses are genuinely deal-blocking versus which are preference, and you can defend that distinction in writing.

What the screen looks for

micro1's process is AI-led and conversational. Expect verifiable specifics about your transactional history — which agreement types, which side of the table, deal volume, what you actually escalated versus signed off on. Then expect follow-ups that go a layer deeper than your first answer: name the clause, state the fallback position, explain why. The screen is trying to separate lawyers who negotiated tech contracts daily from those who supervised someone who did. It also probes evaluation judgment: whether you can score a model's output on defined criteria instead of rewriting it in your own voice, and whether you can articulate the difference between an error and a stylistic disagreement.

Logistics

  • Contractor engagement, part-time, fully remote and largely asynchronous.
  • Typical commitments observed in the 10–20 hrs/week range, with task batches rather than fixed shifts.
  • Occasional synchronous calibration calls with research and product teams, usually US-hours-friendly.
  • Rates observed at $90–150/hr; placement within the band tracks seniority and depth of tech-transactions work, and is not guaranteed.
  • You will need to be comfortable working inside annotation tooling and following written guidelines that change between batches.