About the Role
We are seeking a Robotics Data Quality Engineer to ensure the trustworthiness of our robotics data across all collection modalities. You will analyze, validate, and develop tooling to address data quality issues, bridging the gap between raw collection and usable training data.
Responsibilities
- Analyze robotics data across modalities to identify quality issues, such as plotting joint velocities, validating camera poses, checking gripper encoder accuracy, and flagging anomalous collection sessions.
- Build automated validation pipelines that run on ingestion to catch problems before data enters the warehouse.
- Design and document data formats and schemas across collection modalities, ensuring consistency, versioning, and understanding.
- Develop data visualization tools and dashboards for team inspection and understanding of data.
- Validate cross-modal temporal alignment, including timestamp synchronization, dropped frame detection, and clock drift across various streams.
- Define quality metrics and thresholds per modality and track data quality trends.
- Catalog edge cases and failure modes into a shared taxonomy for common language around data issues.
- Collaborate with data collection operators to trace quality issues to their root cause.
Requirements
- Bachelor's or Master's degree (or equivalent experience) in robotics, computer science, mechanical engineering, or a related field.
- Strong Python data skills (numpy, pandas, matplotlib or plotly) and comfort working with large, messy datasets.
- Solid understanding of 3D geometry, coordinate frames, and spatial transformations.
- Intuition for physical systems; ability to identify issues from trajectory or joint velocity plots.
- Experience designing or working with structured data formats (protobuf, HDF5, ROS bags, or similar).
Skills
- Python
- numpy
- pandas
- matplotlib
- plotly
- 3D geometry
- Coordinate frames
- Spatial transformations
- Physical systems intuition
- Structured data formats (protobuf, HDF5, ROS bags)
- Robotics data analysis
- Data validation
- Automated pipeline development
- Data visualization
- Temporal alignment validation
- Quality metrics definition
- Teleoperation systems
- Motion capture
- Egocentric data collection
- Signal processing
- Sensor fusion
- Time-series analysis
- Data versioning
- Data lineage tracking
- Schema migration
Location
- Remote
Work Type
- Full-time
Experience Level
- 0→1 environments experience
Education Level
- Bachelor's degree
- Master's degree
About the Company
- At xdof, we’re at an inflection point, with frontier labs racing to build general-purpose robots.
- High-quality training data is the bottleneck in robotics.
- We’re building the foundation behind foundation models, including data collection systems, operational capability, exabyte-scale data warehouse, and software toolchain.
- Our goal is to help partners drive the field of robotics forward.
