Staff Software Engineer, Data-Driven Behaviors

The Planner Reasoning Team builds technology that has an extremely broad impact across the Waymo organization and is the team most directly responsible for the Waymo Driver’s behavior. We are the software engineering team responsible for Waymo's high-level motion planning, encoding Waymo's desired driving behaviors onboard and offboard. Our mission is to build the driving logic needed to deliver a safe, excellent Waymo Driver at scale. We have wide reaching impact into both onboard and off-board system architecture, paving the way for future-looking deep-learning-based explorations. In this team, you'll have an impact on scaling our Waymo Driver's performance and maintaining and improving our excellent safety record as Waymo expands to more cities. 

In this hybrid-based role, you will report to an Engineering Manager.

You will: 

• Lead and develop the data pipelines that power ML solutions for the Waymo driver's decision making and motion planning: spanning data mining from large datasets, refining labeling policies, and shaping the reward or loss functions for ML models
• Quickly prototype ideas to improve data pipelines and conduct model finetuning
• Integrate and deploy ML models on the fleet, conducting end-to-end validation and enabling on-field monitoring
• Collaborate closely with partner teams such as perception, research, simulation, and evaluation
• Mentor engineers in leading adoption of data-primary approaches for solving reasoning problems

You have: 

• PhD, Masters or Bachelors degree in Computer Science, Machine Learning, Robotics, or a related field
• 7+ years of years of software engineering / machine learning experience
• Experience in applied machine learning: data shaping, feature engineering, loss/reward shaping, finetuning and model evaluation
• Eval experience: contributing to scalable eval workflows and building metrics for ML models
• Proficiency in dealing with large scale models and datasets. High quality API design with an eye towards eval and data-driven techniques
• Experience with working and solving design problems that cut across multiple components
• PhD, Masters or Bachelors degree in Computer Science, Machine Learning, Robotics, or a related field

We prefer: 

• Experience with Large Language Models (LLM), Vision Language Models (VLM), deep learning or reinforcement learning. 
• Experience in the autonomous driving domain, including areas like motion planning or perception
• Proficiency in C++, Python, and SQL