Senior / Staff Machine Learning Engineer - Behavior Models for Road Users (Autonomous Driving)

Location: Foster City, California, United States; hybrid workplace

Company

Zoox develops autonomous driving systems through its Autonomy Software department.

What you will do

• Develop deep learning models using imitation learning and reinforcement learning to generate driving plans for human-like road-user agents.
• Design techniques to evaluate driving plans for safety, progress, comfort, and realism.
• Build generative behavior models, including autoregressive and diffusion models, conditioned on scenario intent for targeted stress testing.
• Use large-scale compute, infrastructure, and datasets to advance autonomous driving models.
• Develop metrics and analysis tools to identify errors and measure system improvements.
• Collaborate with Perception, Planning, Simulation, and Validation engineers on autonomous driving.

Requirements

• PhD in computer science or a related field with 4+ years of relevant professional experience, or a master’s degree with 7+ years of relevant professional experience.
• Experience in planning, prediction, reinforcement learning, imitation learning, or generative modeling.
• Experience training and deploying transformer-based model architectures.
• Experience with production machine learning pipelines, including dataset creation, training frameworks, and metrics pipelines.
• Fluency with Python machine learning frameworks and basic understanding of C++.
• Ability to work in a hybrid setup in Foster City, California.

Nice to have

• Top-tier publications in venues such as NeurIPS, ICML, or CVPR.
• Experience with JAX.