As an expert in reinforcement learning, you will be working with the Motion Planning Team on developing a machine learning first approach to motion planning that combines the latest findings in reinforcement learning, imitation learning, motion planning, and robotics. You will work in a fast-paced environment and interact with a wide variety of teams ranging across Research, Perception, ML Foundations, ML Infrastructure, Simulation, etc. The ideal candidate should be well versed in the fundamentals of deep reinforcement learning. Your work will directly contribute to our team's ability to build and deploy state-of-the-art deep learning systems for autonomous driving.Responsibilities:
- Develop state-of-the-art methods that leverage deep reinforcement learning, imitation learning, and large-scale data to develop a machine learning first approach to motion planning.
- Train neural networks on massive volumes of data, and build the necessary metrics and introspection tools to enable rapid iteration
- Be a champion of the scientific method and critical thinking in inventing state-of-the-art deep learning solutions but is also a leader in applying rigorous engineering practices during validation and deployment
- Stay up to date with the latest research and trends in the fields and apply new state-of-the-art solutions.
- Collaborate closely with teams such as Perception, Simulation, Infrastructure, Tooling to drive unified solutions
- Work on challenging, unsolved problems and be comfortable with high ambiguity
- Work in a small, high-velocity team of engineers
Lyft is an equal opportunity/affirmative action employer committed to an inclusive and diverse workplace. All qualified applicants will receive consideration for employment without regards to race, color, religion, sex, sexual orientation, gender identity, national origin, disability status, protected veteran status or any other basis prohibited by law. We also consider qualified applicants with criminal histories consistent with applicable federal, state and local law.
- Experience with learning-based planning approaches like imitation learning and reinforcement learning and state-of-the-art techniques like Transformer architectures, muZero, AlphaStar, and GPT.
- Hands-on experience with machine learning / deep learning frameworks such as PyTorch, Tensorflow.
- Advanced degree (Ph.D. preferred) in Machine Learning, Robotics, CS (or other related fields).
- Strong desire to apply critical thinking to tackle challenging real-world problems, resilience to failure, and a passion for deploying solutions into the product.
- Hands-on experience writing high high-quality code in Python and/or C++.
- (Nice to Have) Experience with robot motion planning techniques like trajectory optimization, sampling-based planning, model predictive control, etc
- (Nice to Have) Experience working on self-driving problems (Perception, Prediction, Mapping, Localization, Planning, Simulation)