What the work involves

Most of the day-to-day is judgment work written down. You review model output that touches video production — generated clips, shot descriptions, edit decision lists, color and audio notes, camera and lighting explanations, post-production workflows, codec and delivery specs — and say precisely what is wrong and why. Some tasks are comparative: two model responses, and you rank them with a written rationale that another expert could audit. Others ask you to write the reference answer yourself, at the level of detail you'd give a second assistant editor who has to hit a delivery deadline.

Expect a mix of:

  • Rating generated video or storyboards against a prompt for shot continuity, motion coherence, lighting logic, and physical plausibility
  • Correcting model claims about NLE workflows, transcode settings, frame rates, color spaces, and delivery specs
  • Writing prompts and reference responses that probe where models fail on real production problems
  • Flagging output that is confidently wrong in ways a non-specialist reviewer would miss

What the platform screens for

micro1's screening is AI-led and conversational: a structured interview that asks about your actual credits and workflows, then follows up on the specifics. Vague answers get probed rather than accepted. The bar is a demonstrable professional track record in video — years in an edit bay, on set, in a color suite, or running post pipelines — plus the ability to articulate craft decisions in clear written English. Prior AI or annotation experience is welcome but not the gate; written precision and consistency across many small judgments matter more.

Logistics

Fully remote and asynchronous. Work is contract, invoiced hourly against approved task time, with volume that fluctuates by project — many contributors treat it as 10–20 hours a week alongside client work rather than a full load. There are usually calibration materials and rubrics to read before your first paid batch, and reviewer feedback in the early weeks. Pay bands are as observed for this listing and depend on specialization, screening outcome, and project; they are not guaranteed.