Strong software engineering skills in Python and at least one systems language.
Experience in building large-scale data pipelines using tools like Spark, Ray, or Beam.
Familiarity with cloud storage and compute platforms such as AWS or GCP.
Knowledge of columnar data formats like Parquet or Arrow.
Experience with time-synchronized sensor or video data.
Bonus: Experience with ML training infrastructure or robotics data.
Responsibilities
Build ingestion pipelines for various robot data types.
Design storage, indexing, and retrieval systems for multimodal datasets.
Create tools for labeling, filtering, and curating datasets.
Manage training data infrastructure, ensuring versioning and reproducibility.
Develop evaluation pipelines linking field performance to data insights.
Collaborate with research and operations to address failures and implement fixes.
Benefits
Opportunity to work on cutting-edge robotics technology.
Collaborative and mission-driven work environment.
Emphasis on continuous learning and skill refinement.
High ownership culture with a focus on impactful contributions.
Flexible work hours, emphasizing results over traditional schedules.
Full Job Description
The Role You will own the data flywheel that trains our robots, from collection on deployed fleets to curated datasets ready for training. Your pipelines turn messy real-world data into the fuel for every policy we ship.
What You'll Do
Build ingestion pipelines for robot logs, teleoperation, video, depth, force/torque, and proprioception
Design storage, indexing, and retrieval for multimodal datasets at scale
Build tools for labeling, filtering, search, and dataset curation
Own training data infrastructure: versioning, lineage, and reproducibility
Build evaluation pipelines that tie field performance back to data
Work with research and operations to close the loop from failure to fix
What You Bring
Strong software engineering in Python and at least one systems language
Experience building large-scale data pipelines (Spark, Ray, Beam, or similar)
Familiarity with cloud storage and compute (AWS, GCP) and columnar formats (Parquet, Arrow)
Experience with time-synchronized sensor or video data
Bonus: ML training infrastructure or robotics data experience
Bias toward action: ship fast, measure, iterate
Values
Pace of learning trumps everything.
Relentless refinement of our craft.
Low ego, high ownership.
All-in on the mission. We work in person, and this isn't a 9-to-5.
