7 Voxelmaps Jobs Hiring Near You

Job Summary:
Voxelmaps is building the data infrastructure layer for Physical AI and humanoid robotics. The Lead Robotics Data Engineer owns the technical soul of the operation, focusing on designing data capture protocols, setting standards for datasets, and ensuring operational efficiency across teams.
Responsibilities:
• Author capture protocols for human task, manipulation, navigation, and interaction datasets
• Define dataset schemas, metadata standards, versioning conventions, and acceptance criteria for every dataset family
• Run 20–100 pilot demonstrations per task family in Austin, review failure modes, and iterate
• Operate egocentric (head/chest-mounted), exocentric multi-camera, motion capture suit, and UMI-style demonstration rigs
• Calibrate and maintain all sensor systems before each capture session
• Implement and refine human teleoperation setups as the program matures
• Convert Austin pilot protocols into operator-ready SOPs for the Juárez scaled capture team
• Train Juárez Supervisor on SOPs, certify the capture process, and approve equipment kits
• Monitor protocol drift and recertify operators as dataset families evolve
• Support training and evaluation of VLMs and VLAs on robot arms and other platforms in Austin
• Close the loop between data collection and model evaluation outcomes — stop collecting low-value data quickly
• Chair weekly protocol review: new task families and failure analysis
• Collaborate with the Data Engineer / MLOps to build ingestion, versioning, metadata integrity, and dataset packaging workflows
• Monitor metadata completeness and time-from-capture-to-packaged-dataset KPIs
• Contribute to partner-grade dataset documentation and train/validation splits
• Present experimental findings internally and to external partners
• Produce technical reports and maintain a clear dataset release log
• Contribute to talks, demos, and video presentations as Voxelmaps grows its external profile
Qualifications:
Required:
• Hands-on robotics experiment design & protocol authorship
• Capture architecture: egocentric, exocentric, mocap, UMI-style
• Python — strong, production-quality scripting
• ROS / ROS2 — practical experience
• Dataset standards: metadata schemas, versioning, QA acceptance criteria
• Ability to translate protocols into Juárez operator-ready SOPs
• Cross-functional communication: engineers, operators, partners
• Self-directed; owns outcomes without close supervision
Preferred:
• Responsible for interviewing, selecting, and building the robotics division. This division will consist of two teams (experimentation research and definition team and data acquisition team); assumes full ownership of team performance, execution, and overall success
• MSc in Robotics, CS, Mechanical Engineering, or related (strong industry background accepted)
• Imitation learning / robot learning pipeline experience
• Human teleoperation system operation
• Familiarity with open-source VLM/VLA models (LLaVA, OpenVLA, RT-X)
• Motion capture suit calibration & management
• Experience with humanoid robot platforms
• Prior experience writing training or QA documentation for operators
Company:
Welcome to Voxelmaps: Creating Data for Robots Founded in 2017, the company is headquartered in Austin, USA, with a team of 201-500 employees. The company is currently Growth Stage.