I Joined Google (again)
This Spring I joined Google for the second time. I’m working on core ML infrastructure, specifically ML frameworks for Keras, JAX, and Gemini.
Not everyone is aware that Waymo is a completely separate sister company to Google. Going between the companies is an external hire in both directions.
Since joining Waymo four years ago, I worked with a ton of excellent people on a lot of different aspects of ML infra, including quantization-aware training, ML training frameworks, and global-scale automation of TPU training and simulation pipelines. It was always super rewarding walking around the streets of SF encountering a Waymo doing a specific behavior and thinking to myself “I worked on that model!!” (at least in very small part 🙂)
One aspect of working at Waymo I had not experienced before was being part of a new transformational technology that required a big shift in societal expectations. Living in San Francisco the entire time, I witnessed first-hand a clear arc from Waymo’s initial testing in 2021, where the cars—initially very rare—were viewed with surprise and suspicion, moving methodically toward their current state of wider adoption. I attended the CPUC hearing in 2023 where Waymo got its initial authorization to carry members of the public in SF, and had my photo taken while getting into a Waymo make it into the news in a robotaxi hit piece.
Joining Google now, I’m super excited to get exposed to a very diverse array of ML challenges and use cases. I’m also really excited that part of my team’s work is open-source first, so I’m very excited to help the open source community in a larger capacity again and share more about what I work on.