Video Annotation Expert (AI Training – Robotics)

Job Summary

Mercor is seeking Video Annotation Experts to help train AI systems and robotics by analyzing first-person (egocentric) videos of everyday household tasks.

In this role, you will break down real-world human actions into structured, labeled segments, enabling AI systems to understand and replicate tasks such as cleaning, organizing, and handling objects.

Key Responsibilities

• Watch first-person videos of people performing household tasks.
• Segment videos into precise action intervals with frame-level accuracy.
• Label each action with timestamps (start/end).
• Write clear, natural language descriptions for each action.
• Identify which hand(s) are used for each action.
• Screen videos for quality before annotation.
• Follow detailed annotation guidelines for consistency.

Example Tasks

• Segment actions like: "pick up mug → walk to counter → place mug on surface"
• Label actions with accurate timestamps and contextual descriptions.

Requirements

• Strong attention to detail and precision.
• Ability to interpret guidelines and apply judgment in ambiguous situations.
• Clear, concise English writing skills (native or near-native).
• Comfortable performing repetitive, detail-oriented tasks.
• Basic computer proficiency and ability to learn new tools quickly.

Preferred Qualifications

• Experience with video editing, annotation, transcription, or data labeling.
• Familiarity with media workflows or content analysis.

Eligibility

Applicants must be based in:

• United States (excluding CA, NY, CT, WA, DC)
• Canada

Contract & Payment

• Employment Type: Contract (Freelance)
• Location: Remote (USA & Canada only)
• Pay: $16 per hour
• Payments are made weekly via Stripe or Wise.

Application Process

• Complete a short AI interview (~20 minutes)
• Take a screening quiz (~15–20 minutes)
• Selected candidates may receive offers within 1 day