Egocentric Video Data Reviewer

Review first-person video data and make precise, consistent quality decisions across a series of short-term annotation projects. The first assignment focuses on validating action-segment boundaries, with additional egocentric-video projects following during the engagement. Key Responsibilities

• Review egocentric video data and complete structured quality-review tasks using project-specific guidelines.
• For the first project, assess action segments identified within each video.
• Determine whether each segment''s start and end points accurately match the action performed.
• Adjust or flag incorrect boundaries according to the provided guidelines.
• Apply consistent judgment across a high volume of short action segments.
• Complete assigned videos within the project timeline.

• Review 50 egocentric videos, each approximately 1 to 2 minutes long.
• Review an average of approximately 38 action segments per video.
• Allow about 30 minutes of review time per video.
• Receive detailed instructions at the start of the project.

• Excellent attention to detail and visual comprehension.
• Ability to distinguish carefully between closely related actions and moments in video.
• Comfort with structured, repetitive review work while maintaining accuracy.
• Ability to learn and consistently apply detailed annotation guidelines quickly.
• Strong written English communication skills.
• Prior experience with video annotation, data labeling, computer vision datasets, or egocentric video is helpful but not required.

• Remote, hourly engagement.
• Five openings are available.
• Approximately two weeks of work across three short-term projects.
• Each project is expected to last approximately 3 days to 1 week.
• The first project is expected to begin Monday morning, August 17, Pacific Time.

$20 to $25 per hour.