What the work actually is

You watch recorded sessions of robotic arms doing physical tasks — picking up a mug, stacking blocks, opening a drawer, pouring — and mark what happened. That means segmenting the video by event, labelling object interactions (contact, grasp, release, slip), tracking spatial changes, and grading whether the task was completed as specified. When the footage is ambiguous — the gripper is occluded at the moment of contact, motion blur hides a drop, the camera angle makes it impossible to tell whether the object cleared the lip — you flag it and write a short note explaining what you can and cannot see. That written note matters as much as the label.

The reason a VFX background is valued here is specific: you already read motion as data. You know what a trajectory looks like when it's been retimed, what real contact looks like versus a near-miss at a bad angle, how occlusion resolves frame by frame, and when something in a plate is physically implausible. None of the pipeline work transfers — there is no modelling, no lookdev, no compositing. What transfers is your eye and your vocabulary for describing what it saw.

What the screen looks for

micro1's screening is AI-led and conversational. Expect it to probe your actual production history — what you animated or comped, on what software, at what scale — and then push on motion judgment: describe an occlusion, explain how you'd determine contact from a partial view, reason about physical plausibility. It also tests guideline discipline. The failure mode this project cares about is an experienced artist who substitutes their own aesthetic judgment for the customer's written rubric. Expect at least one scenario where the guideline says something you'd personally disagree with.

Logistics

  • Contractor, fully remote; open to candidates in North America, Latin America, Europe, the UK, and Oceania.
  • Asynchronous batch work with deadlines rather than fixed shifts; volume varies by project phase.
  • Observed rates for this listing run $50–90/hr, typically scaled to background depth and review-tier responsibility. Rates are as observed, not guaranteed.
  • No prior AI, robotics, or annotation experience required — the annotation tool and guidelines are provided.