Senior Machine Learning Engineer (Infra), Driver Understanding and Evaluation

The DUE Machine Learning team will build and operate scalable machine learning and data systems, simulation workflow and insight tools, improve and speed up the evaluation and onboard developer journeys. It will combine expert human judgements and advanced machine learning models to deliver training and evaluation data for hundreds of metrics and components that make up the Waymo driver. We are looking for researchers and software engineers who are passionate about developing machine learning techniques for the Evaluation systems on our autonomous vehicles, and have an incessant drive to improve the performance of our technology stack.

You will:

• Build scalable systems for training and fine-tuning large-scale models to evaluate interesting driving behaviors.
• Work at the intersection of data engineering, model development, and simulation Provide guidance on architectural decisions and technical directions. Own large, complex systems, driving architectures that meet technical and business objectives.
• Contribute to the production and optimization of machine learning models aiming to assess Waymo’s expansive fleet of vehicles that cumulatively travel millions of miles.
• Design and scale large distributed systems covering the ML lifecycle, supporting planet-scale dataset generation, model training, and evaluation.
• Collaborate cross-functionally to derive performance and system-level requirements for large ML systems. Translate product/business goals into measurable technical deliverables, ensuring system component alignment.

• M.S. or Ph.D. degree Computer Science, Machine Learning, Artificial Intelligence, or a related technical field, or equivalent practical experience.
• 5+ years in machine learning infrastructure such as developing, designing, scaling, training, deploying, and optimizing large-scale machine learning systems from data to model.
• A history of contributions to machine learning tooling and frameworks e.g. PyTorch, Jax, Tensorflow, Ray, or similar. The candidate should understand both the user facing API and the internal workings. 
• Strong expertise in distributed training techniques, including gradient sharding and optimization strategies for scaling large models across ML accelerator profiling tools to uncover performance bottlenecks.

Nice to have

• 7+ years in machine learning infrastructure such as developing, designing, scaling, training, deploying, and optimizing large-scale machine learning systems from data to model.
• Experience in the autonomous vehicles domain, robotics, or complex simulation environments.
• Familiarity with large-scale simulation platforms and their integration with ML training workflows.