The company is looking to solve the problem of introspecting machine learning-based motion planners, specifically developing techniques and tooling to understand and debug motion planning behavior.
Requirements
- Proficiency in Python and ML libraries (PyTorch, NumPy, Jax)
- Strong understanding of Imitation and Reinforcement Learning concepts and motion planning search techniques
- Familiarity with common learned model introspection techniques
- Familiarity with autonomous vehicles or robotics
- Solid understanding of LLM concepts
Responsibilities
- Lead new initiatives to introspect the output of machine learning motion planning models. You will both utilize existing cutting edge techniques as well as develop novel introspection techniques.
- Design new architecture and implement new tools used to analyze machine learning model behavior.
- Collaborate with engineers on Perception, Prediction, Planning, Machine Learning Model, and Visualization teams to enable development of behavior improvements.
Other
- MS or PhD in Computer Science, Machine Learning, or a related field
- 5+ years of industry experience in Machine Learning–experience building and maintaining ML training pipelines, including data preprocessing, model training, and evaluation
- Publications in top-tier conferences (CVPR, ICCV, RSS, ICRA)
