Responsibilities
- Own the robot-learning pipeline end to end. Establish and maintain a trusted workflow spanning robot data collection, data ingestion and compilation, post-training, checkpoint generation, inference, and real-robot evaluation.
- Define the supported golden path. Maintain known-good combinations of code, datasets, configurations, checkpoints, robot software, hardware settings, task stations, and evaluation procedures.
- Build automated validation at every interface. Develop checks for timestamp synchronization, sensor and action integrity, episode completeness, schema compatibility, dataset migrations, dataloader outputs, preprocessing behavior, model inputs, and configuration correctness.
- Create end-to-end regression tests. Build representative smoke tests that exercise data compilation, training, checkpoint loading, inference, replay or simulation, and real-robot execution. Develop small-scale overfit and canary experiments that catch correctness regressions before expensive training runs begin.
- Ensure training and inference consistency. Identify and prevent discrepancies in image processing, sensor normalization, temporal context, action representation, model configuration, and other transformations used across training and deployment.
- Make failures observable and diagnosable. Build instrumentation and debugging tools that help determine whether a performance regression originated in data, model code, infrastructure, inference, robot software, hardware configuration, the physical environment, or evaluation execution.
- Improve real-robot evaluation reliability. Partner with researchers and robot operations to establish stable benchmark stations, reference baselines, clear rubrics, repeatable trial protocols, operator procedures, and tracking of environmental variables that affect performance.
- Lead cross-functional root-cause investigations. Drive ambiguous failures to resolution across Research, Data Infrastructure, Model Infrastructure, Software, and Robot Operations. Turn incidents and regressions into durable tests, monitors, documentation, and interface contracts.
- Establish release and compatibility standards. Define the validation required before changes to robot software, data systems, training code, or inference systems become part of the supported research workflow.
- Measure and improve research reliability. Track pipeline success rates, reproducibility, regression frequency, time to root cause, benchmark stability, and other metrics that reflect the health of the research workflow.
Requirements
- A strong track record owning, integrating, or debugging complex ML, robotics, autonomy, or other sensor-rich systems across multiple layers.
- Excellent software engineering skills, including strong Python, maintainable system design, automated testing, CI, code review, and production-quality debugging practices.
- Hands-on experience with PyTorch or a comparable ML framework. You should be able to train or fine-tune models, inspect datasets and batches, interpret losses and outputs, load checkpoints, and debug inference behavior.
- Experience building or operating ML workflows that span data processing, training, evaluation, and deployment - not only one isolated part of the stack.
- Strong systems-debugging ability. You can take an ambiguous report such as "robot performance dropped" and turn it into a structured investigation with controlled experiments, measurements, and a defensible root cause.
- Familiarity with data-pipeline correctness, including schemas, versioning, temporal data, migrations, provenance, validation, and reproducibility.
- Experience with Linux, containers, and modern compute environments such as Kubernetes, Slurm, or distributed GPU clusters.
- Strong attention to detail and a willingness to investigate silent failures that do not produce obvious exceptions or crashes.
- Ability to work effectively across research, infrastructure, software, and operations teams and to drive decisions without relying on formal managerial authority.
- Clear written communication, including the ability to document interfaces, test plans, root-cause analyses, release criteria, and operational procedures.
- Comfort working directly with physical systems and dealing with the variability and ambiguity inherent in real-world robotics.
- A degree in Computer Science, Robotics, Electrical Engineering, or a related discipline - or equivalent practical experience. A PhD and publication record are not required.
Nice to Have
- Experience in robotics, autonomous driving, drones, industrial automation, warehouse automation, or another domain that combines learned models with physical systems.
- Familiarity with robot data collection, teleoperation, camera and sensor systems, calibration, proprioception, force/torque sensing, or action synchronization.
- Experience with ROS or ROS2, real-time robot systems, simulation, replay infrastructure, or hardware-in-the-loop testing.
- Experience with imitation learning, behavior cloning, robot post-training, reinforcement learning, video models, multimodal models, or learned control policies.
- Experience building golden datasets, model-quality regression suites, data contracts, or automated dataset validation.
- Familiarity with columnar datasets and data infrastructure, including formats such as Parquet and systems for indexing, querying, or migrating large datasets.
- Experience debugging distributed training, GPU environments, checkpointing, networking, storage, Kubernetes, Slurm, or InfiniBand-related failures.
- Experience designing statistically sound evaluation protocols for systems with noisy or stochastic performance.
- Prior experience in ML platform engineering, research infrastructure, autonomy validation, release engineering, SRE, or systems integration.
