RoboticsDataQualityEngineer

At xdof, we’re at an inflection point. Frontier labs are racing to build general-purpose robots, and high-quality training data is the bottleneck. We’re building the foundation behind the foundation models – the data collection systems, operational capability, exabyte-scale data warehouse, and software toolchain – to help our partners drive the field forward. The models are only as good as the data. We’re looking for a Robotics Data Quality Engineer to be the person who knows whether our data is trustworthy, across every modality we collect. You’ll analyze, validate, and build tooling around data from teleoperation on real hardware, egocentric capture, UMI-style grippers, and more. If something is wrong with the data, you’re the first to catch it and the one who helps us fix the process. What You’ll Do Data quality engineers are the bridge between raw collection and usable training data. Sample projects include: - analyzing robotics data across modalities to identify quality issues: plotting joint velocities, validating camera poses, checking gripper encoder accuracy, and flagging anomalous collection sessions - building automated validation pipelines that run on ingestion and catch problems before data enters the warehouse - designing and documenting data formats and schemas across collection modalities, ensuring they are consistent, versioned, and well-understood by partners and internal researchers - building data visualization tools and dashboards so the broader team can inspect and understand the data without writing custom scripts - validating cross-modal temporal alignment, including timestamp synchronization, dropped frame detection, and clock drift across camera, joint, and gripper streams - defining quality metrics and thresholds per modality and tracking whether data quality is improving or degrading as collection scales - cataloging edge cases and failure modes into a shared taxonomy so the team has a common language for data issues - working closely wit