About Role
We're building a Vision-Language-Action (VLA) dataset to train next-generation humanoid robots. Your role is to record short, first-person videos of everyday household activities using your SmartPhone (1080p, head-mounted view) .
Each clip (1–2 minutes) should show a natural human completing one task, e.g., folding laundry, loading dishes, wiping a table, or pouring water, from a consistent first-person perspective .
Responsibilities
- Record clear, stable first-person household activity videos following provided task scripts.
- Maintain proper lighting, camera angle, and framing.
- Upload raw videos to a secure drive following naming and organization guidelines.
- Ensure diversity in environments (different surfaces, rooms, lighting).
- Maintain data quality and consistency across all recordings.
Requirements
Access to a recent Smartphone (1080p @ 30 FPS) or equivalent device.Ability to mount the camera at head level (head strap or fixed mount).Stable indoor environment suitable for recording short household tasks.Basic English comprehension (for instructions and task labeling).Reliable internet connection for video uploads.Example Tasks
Folding a towelLoading and unloading a dishwasherOrganizing desk itemsPouring water into a glassWiping a kitchen counterSkills Required
Smartphone 1080p