Remote Household Data Specialist for Video Capture

Capture high-quality video demonstrations of everyday household tasks using a UMI gripper wearable device, producing training data that teaches personal robots to operate in real homes. This contractor role is remote and Vietnam-based, and requires precise, repeatable recordings that follow detailed task specifications and reviewer feedback.

• Record and submit specification-compliant video demonstrations of household tasks using a UMI gripper wearable device.
• Follow detailed, task-specific instructions to ensure every recording meets project standards.
• Set up and operate recording equipment and the UMI gripper so hand movements accurately mimic robot behavior.
• Review feedback from assigned reviewers, iterate on technique, and improve data quality over time.
• Communicate clearly and proactively with the project team, both in writing and verbally.
• Maintain a consistent weekly workflow to meet minimum approved data requirements.

Required

• Experience with or ability to use UMI-grippers and related wearable devices.
• Proven skill in capturing high-quality video, with strong attention to detail and ability to follow complex specifications precisely.
• Manual dexterity and comfort performing precise physical tasks in household environments.
• Comfort operating and wearing specialized devices and recording equipment, including prolonged use of a VR headset as part of regular work.
• Ability to receive and incorporate iterative feedback to meet quality benchmarks.
• Strong written and verbal communication skills.
• Reliable access to a fully equipped Philippine household environment, including kitchen and common living spaces and common items.
• Stable, high-speed internet capable of uploading large video files from your home.

Preferred

• Curiosity about personal robotics, AI, and emerging home technologies.
• Prior experience in data collection, usability studies, or hands-on, precision-oriented projects.
• Patience for repetitive, detail-focused work and flexibility as project needs evolve.

• Engagement type, contractor.
• Remote, Vietnam-based candidates only.
• Independent contractor, pay-per-task model where you are paid per hour of approved data that meets project specifications.
• Maintain a consistent weekly workflow to satisfy minimum approved data requirements.
• Selection and onboarding are time-sensitive, candidates should be prepared to start quickly after selection.

• Rate range, 15 to 30 USD per hour of approved data.
• Pay-Per-Task structure, compensation is for approved data hours only.
• Pilot phase, an initial rate per approved hour applies until calibration and quality benchmarks are met.
• Post-calibration, a higher per-approved-hour rate becomes available after consistently meeting quality standards.
• Planning benchmark, expect roughly 2 hours of review work per 1 hour of approved data early on, yielding an effective fixed hourly rate during the pilot, with typical improvement as you gain experience.

• Applicants must be able to work remotely while based in Vietnam.
• Must have reliable home access to the household environment described above, and stable high-speed internet for large file uploads.
• Be comfortable with prolonged VR headset use when required by tasks.
• Roles are often filled quickly, typically within 48 hours, so applicants should be ready to begin their first task as soon as they are selected.

• Submit your application and confirm your availability to start on short notice.
• If selected, you will receive task instructions, device setup guidance, and recording specifications.
• Submit initial video data for review, receive feedback from reviewers, and iterate until calibration and quality benchmarks are met.
• Once calibration is achieved, you become eligible for the improved per-approved-hour rate and ongoing task assignments.

micro1 is an AI data lab focused on creating high-quality training data and evaluations for frontier AI models. This project contributes to training personal robotics by turning real-world human demonstrations into machine learning data.