Robotics Data Operator

Faraday Future•El Segundo, CA

2d•Onsite

About The Position

As a core member of the Robotics Data Factory, you will be responsible for collecting high-fidelity human behavior and robot teleoperation data. This data is essential for training the next generation of Vision-Language-Action (VLA) models, helping robots learn to perceive and interact with the physical world by capturing real-world human interactions and precision robotic movements.

Requirements

0–1 years of experience.
Currently pursuing or recently completed a Bachelor's or Master's degree in Computer Science or a related field.
Extreme Attention to Detail: Ability to strictly follow complex technical guidelines regarding framing, motion rhythm, and lighting environments.
Technical Proficiency: Comfortable using smartphones, VR hardware (Meta Quest/HTC Vive), and various digital accessories. Basic understanding of video recording parameters and calibration.
Action Consistency: Ability to perform tasks naturally while maintaining the specific "robotic-friendly" pacing required for imitation learning.

Nice To Haves

Experience in Python, Linux command line.
Familiarity with ROS2, LeRobot, or AI data labeling workflows.
Strong interest in Embodied AI and robotics.

Responsibilities

Egocentric Data Collection: Wear designated head-mounted recording devices (e.g., smartphone rigs, GoPro, or head-mounted VR kits) to capture daily or industrial tasks from a first-person perspective in diverse environments (home, retail, or manufacturing).
VR Teleoperation: Operate robotic systems via VR interfaces to perform complex manipulation tasks. Ensure high-quality capture of multi-modal data, including joint states, torques, and gripper status, while maintaining low-latency execution.
Technical Specification Compliance: Maintain strict recording standards (1080p, 30FPS, horizontal orientation, FOV >130°). Ensure accurate extraction of camera intrinsic parameters and time-synchronization across all sensors.
Quality Assurance & Pre-processing: Conduct initial self-inspections to filter out clips with excessive shaking, occlusions, or focus failures. Ensure data is correctly formatted (e.g., MCAP, HDF5) and uploaded according to protocol.