What the work actually involves

You will spend most of your time in two modes. The first is production: drafting or redlining clauses in realistic tech-transaction scenarios — a limitation of liability that needs to survive a vendor's pushback, a DPA schedule that has to reflect actual sub-processor reality, an NDA residuals clause a counterparty slipped in. Some tasks are structured as simulated negotiations, where you play one side across several turns and your markup is captured as training signal. The second mode is evaluation: reading a model's redline or memo and judging it — was the risk correctly identified, was the fallback position commercially sensible, did it hallucinate a defined term or misstate what the indemnity actually covers? Feedback needs to be specific enough that a researcher who is not a lawyer can act on it.

A meaningful slice of the role is rubric work. You will be asked to turn tacit judgment into written grading criteria — what separates a strong response from a passable one on a change-of-control clause, and how two attorneys reviewing the same output land on the same score. Expect to iterate on guidelines with product and research staff and to defend your calls when another reviewer disagrees.

What the screen looks for

micro1's process is AI-led and moves quickly. It verifies the J.D. and active bar admission, then probes transactional depth: which agreements you have actually negotiated, from which side of the table, and what positions you took. Generic litigation or compliance experience does not substitute. Screeners follow up on specifics — naming a clause is not the same as explaining the commercial reason a client accepted it. They also test whether you can write criticism that is concrete and non-hedged, since legal instinct expressed as "this feels off" is unusable as training data.

Logistics

  • Remote, contractor, part-time — typically a batch of tasks with a turnaround window rather than fixed hours.
  • Largely asynchronous, with occasional calibration calls with reviewers or research staff.
  • Volume fluctuates with project cycles; this is a supplement to practice, not a replacement for it.
  • Pay observed in the $80–200/hr range, set by experience and task type, and not guaranteed for any given engagement.