Nate Harada Machine Learning in Real Life



I’m Nate, currently working with the perception team at Waymo to develop self-driving vehicles. I used to work at Cruise, where I built systems to detect objects, respond to emergency vehicles, and track things over time. I also worked at Fitbit in the R&D department, where I researched how we can apply advanced AI to sensors and health data to make people healthier. In grad school I worked in Zeeshan Syed’s lab, investigating how deep learning can improve our understanding of heart diseases such as atrial fibrillation.

I also am the creator and maintainer of the open source tool Moonshine, downloadable pre-trained ML models for remote sensing and satellite data. It’s my hope that by reducing labeling costs via pre-training, climate researchers and policy makers can better do their jobs.

My interests lie at the intersection of machine learning and the real world: I most enjoy work on problems related to images, sensor data, audio, or other physical data. I believe that as computation and sensors are further woven into the fabric of our lives, we need smarter interaction models to help us make sense of it all.

If you need help figuring out machine learning in real life, I’m happy to help. I ❤️ pro bono work with non-profits – if you have a good cause that needs proven ML expertise and experience, especially anything related to solving the climate crisis, I’d love to chat.