Principal Software Engineer - Waymo Driver Post Training

The Perception team at Waymo builds technology that powers the Waymo Driver. Our models allow the Waymo Driver to perceive the world around it, make the right decision for every situation, and deliver people safely to their destinations. We conduct research to address real-world problems and collaborate with research teams at Alphabet. We have access to millions of miles of driving data from a diverse set of sensors, allowing ML practitioners like you to develop multi-modal models and techniques at scale.

In this hybrid role, you will report to our Senior Director of Engineering, Perception.

You will:

• Work closely with Onboard model engineers on rapid prototyping and experiments, to verify model design and training recipes, before decision to productionize

• Whenever needed, prototype the necessary infrastructure (data, eval) to enable model experimentation

• Drive post-training technology roadmap (e.g., RL techniques) and training recipe roadmap

• Drive and contribute to the production rollout of major model architecture and recipe evolutions, to solve a variety of behavior contexts

• Build and facilitate strong collaborations between Onboard and research teams across Waymo and Alphabet

• Work closely with Onboard and collaborators to build and continuously evolve the post-training recipe roadmap, with milestones towards strong capability goals and continuous production impact
  
  

You have:

• Remaining technical and hands-on, either currently as an individual contributor, or leading a small team

• Actively writes code, reviews code, and runs experiments

• Hands-on experience in post-training techniques, e.g., RL and SFT

• Experience in at least one of Robotics, VLA, VLM. Strong in both theoretical and practical understanding in model architectures and training techniques; built SOTA models; built / extremely familiar with the entire model development process

• Production driven experience and mindset - problem driven, strong ownership on quality and timely delivery, rigor in evaluation

• Organizational awareness, extremely collaborative, strong communication, focus on value add to the company and to the team

We prefer:

• Experienced at model optimization, model-compute codesign

• Experienced at driving cross-org efforts in transitioning large systems through major evolutions

• Experienced at modern ML data, training, and inference systems for large models

• Strong track record coaching and growing engineers and leaders

• Domain experience in either robotics or AV