What the work involves

You will be doing the thing you already do — marking up an MSA, pushing back on an uncapped indemnity, catching a DPA that quietly grants onward transfer rights — except the counterparty is a model, and your redlines become training signal. Tasks typically take one of three shapes: a simulated negotiation where you play one side across several turns of markup; a review pass where an AI has redlined a contract and you grade the output clause by clause; or a framework task where you write the rubric another reviewer will use to score model responses on a class of contract problems.

The grading work is where most contractors either excel or wash out. Saying a model's limitation-of-liability edit is "wrong" earns nothing. Saying it carved consequential damages out of the cap but left the IP indemnity inside it, that this is a client-side position no vendor would accept, and that the model failed to flag the inconsistency — that is the artifact the research team needs. Feedback has to be legible to someone who is not a lawyer and reproducible by someone who is.

What the platform screens for

micro1 runs an AI-led interview before any human review. Expect it to probe your actual deal history: which agreements you owned end to end, what your standard fallback positions were, how you handled a specific negotiation that went sideways. The screen follows up on vague answers, so "I negotiated a lot of SaaS agreements" will be met with a request for the clause you fought over most and why. It also tests whether you can articulate a grading standard rather than just a legal opinion — the difference between "I'd redline that" and "here is the criterion under which that redline is either correct or defensible."

Logistics

  • Fully remote, contractor engagement, asynchronous — no fixed hours or standing calls.
  • Part-time by design; most contributors fit tasks around a full-time role or practice.
  • Compensation is per completed task that passes review, averaging around 3.5 hours of handling time. The $90–130/hr band is an observed effective rate, not a guarantee; it moves with your throughput and your sign-off rate.
  • Occasional async collaboration with product and research teams on guidelines and edge cases.