Github

See my github different active and (sadly) defunct code that I've worked on. I am actively working on some machine learning code, which encapsulates some of the techniques that I've learned. The code is in active development, as is the documentation, so buyer beware.

Autonomous Vehicles

From 2015 until 2023 I worked on autonomous vehicles at NVIDIA. I was the first hire on a team whose mission was to validate end-to-end learning for the AV task. We were successful, moving from "research" and into "production" in under two years. We held a live demo open to public walk-ins during CES 2017, less than 2 years into the project It was covered by various media, such as The Verge and TechCrunch.

Within five years, we achieved distances of around 500km between failures along United States highways, driving in all weather and lighting conditions: night, day, rain, and snow. We also covered multiple states along the East Coast in single test drives. A video with clips of a multi-state drive is on the NVIDIA blog: "Researching and Developing an Autonomous Vehicle Lane-Following System".
We published a pair of interesting arxiv papers as well.
End to End Learning for Self-Driving Cars in 2016
Explaining How a Deep Neural Network Trained with End-to-End Learning Steers a Car in 2017

Autonomous Drones

In 2024, I worked with a small startup company creating autonomous guidance for drones. I created the data collection, labelling, and training pipelines, and we were able to demonstrate real-time autonomy using only a Raspberry Pi Zero for on-board compute and a Picam sensor as input.