Data Type
Task Types
Subject Matter / Industry
Share link
4 ground-truth units per problem
200,000 ground-truth units
Candidates should have proven experience with video annotation, data labeling, or similar data-centric projects, along with exceptional visual-data accuracy, attention to detail, and strong written and verbal communication. They should be self-organized, capable of independent time management, familiar with annotation tools, and able to collaborate on ambiguity resolution and quality reviews. Experience in AI training, machine learning data preparation, or robotics is strongly preferred, but prior AI experience is not required. Contributors will review videos of robotic arms performing assigned tasks, identify key actions and outcomes, and apply detailed grading guidelines to tag and annotate events. They will use annotation platforms to record structured observations, document findings, submit datasets according to project standards and milestones, communicate progress and challenges, collaborate with trainers and peers, and address feedback during ongoing quality reviews.