What the work actually looks like

Video annotation at this pay band is not box-drawing piecework. Contributors are typically brought in for tasks where judgment is the bottleneck: deciding exactly which frame an action begins on, keeping an object's identity consistent across occlusion and re-entry, writing temporally grounded descriptions that a model can be scored against, or reviewing another annotator's output and deciding whether a disagreement is an error or a guideline gap. Some projects are pure labeling; others are evaluation — you watch two model-generated video captions or two sets of predicted tracks and judge which is better, with a written rationale that has to survive a reviewer reading it cold.

Expect to work inside a client's tool rather than one of your choosing: CVAT, Label Studio, SuperAnnotate, V7, or an in-house web annotator. Guidelines are long, occasionally self-contradictory, and get revised mid-project. A meaningful share of the job is reading the spec closely, applying it consistently across hundreds of clips, and flagging the edge cases that the spec doesn't cover instead of quietly inventing a rule.

What the screen looks for

micro1 runs an AI-led interview before any project match. It probes for concrete history — which tools, which annotation types, what volume, what quality metrics you were held to — and it follows up when an answer stays abstract. It also tests calibration: whether you can articulate why one label is defensible and another is not, and whether you distinguish "the guideline is ambiguous here" from "I disagreed with the guideline." Claims about throughput or IAA scores should be numbers you can actually name.

Logistics

  • Fully remote and largely asynchronous; deadlines are per-batch rather than per-hour.
  • Reliable broadband and a display large enough for frame-by-frame scrubbing are practical necessities.
  • Engagements are typically part-time and bursty — 10–25 hours a week during an active batch, then gaps.
  • Pay is stated as observed across similar listings, not guaranteed; rate depends on project complexity and the reviewer tier you qualify for.
  • Written English strong enough for description and rationale tasks is assumed throughout.