About the role
You produce the gold-standard segmentation and labels that everything downstream depends on. Working frame by frame, you segment it and write a exact, per-hand command for each task - to a 99%+ accuracy bar. You also help author the SOP, calibrate new annotators, and verify others' work as a ground-truth specialist reviewer.
What you'll do
Segment egocentric videos into frame-accurate action boundaries Write action labels using a controlled guidance driven vocabulary Hold a 99%+ accuracy bar Act as a ground-truth reviewer: audit and correct other annotators' segments and labels Sharpen the annotation guidelines and edge-case rules as the specification evolves What we're looking for - must have
Hands-on experience with video annotating for robotics, embodied AI - ideally at or for a leading robotics company or project A real feel for frame-level precision: you can explain what a 99% (versus 95%)
accuracy bar means in practice, and you work to it Outstanding written English - your labels are commands a model learns from, so clarity and consistency are non-negotiable Meticulous, consistent, and comfortable with focused, repetitive precision work over long sessions Experience with professional annotation tooling and frame-by-frame video review Experience authoring or following detailed annotation SOPs and rubrics Nice to have
Robotics, mechanical/mechatronics engineering, or computer-vision background Engineering background from a tier 1 institution or worked at an AI/robotics startup Why it matters
Your ground truth becomes the benchmark and the model is measured against it. This is the role that sets the standard.