The work
- Review 2D and 3D human animation sequences for natural movement and physical plausibility in embodied AI evaluation and benchmarking.
- Assess foot placement, body balance, center-of-mass movement, and anatomical joint positioning.
- Identify foot sliding, joint jitter, joint pops, body intersections, floating body parts, and unnatural poses.
- Judge each animation clip as a whole and apply standardized ratings from 1 to 5.
- Rate the severity, frequency, and persistence of observed motion issues across multiple sequences.
What it takes
- The project is open worldwide.
- Fluency in English is required.
- Strong visual observation, close attention to detail, and a basic understanding of human anatomy, joint movement, and physical motion are required.
- You must distinguish natural from unnatural movement, recognize animation artifacts and physical inconsistencies, follow structured evaluation guidelines, and maintain concentration during repeated visual review.
- Helpful backgrounds include animation, motion capture, biomechanics, game testing, visual annotation, motion-quality evaluation, or rubric-based scoring.
- Accurate, repeatable judgment and clear communication of ratings are important for this project-based work.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training is the human work behind systems that learn from examples, including reviewing video and rating model outputs. People with relevant skills are paid to make careful judgments that help improve how AI systems understand and respond to the world.