perit.ai
MultimediaVideo

Video Scene Semantic Tagger

Perform temporal scene segmentation, action recognition labelling, and multi-modal event annotation on video datasets for frontier AI training in computer vision and video understanding.

Rate
$45–$95 / hr
Placed this month
156
Commitment
Flexible, 10–30 hours per week
Responsibilities

What you'll do

  • Segment video into semantically coherent scenes with precise timestamps.
  • Label human actions, object interactions, and spatial relationships.
  • Annotate multi-modal events combining visual, audio, and text cues.
  • Verify AI-generated video captions and descriptions for accuracy.
Requirements

What we need

  • Experience in video annotation, film production, or computer vision labelling.
  • Strong spatial and temporal reasoning skills.
  • Familiarity with annotation tools (CVAT, Label Studio, or equivalent).
  • Attention to detail on frame-level and scene-level tasks.
Ready to apply

15 minutes. One interview. Verified for life.

Complete one domain interview and you'll be matched to paid research work — including this role — within a week.