Aboutupdated June 2021
I work at the intersection of deep learning and computer architecture. I think ML algorithms should inspire silicon architectures, and vice-versa. To that end, I currently work at Apple on the Neural Engine project.
Previously, I graduated from Stanford with a BS and MS in electrical engineering. My coursework centered on AI / ML algorithms, with a special focus on computer vision and graphics. In my summer internships, I spent two summers at NVIDIA in the Deep Learning Architecture group, writing CUDA and learning about GPU microarch to train huge models at maximum speed. I also spent a very exciting summer at Built Robotics, where I helped design an autonomous vehicle's perception system from the ground up.
Now, I'm interested in on-device inference, where efficiency is king. This is a multi-pronged effort, built on bleeding-edge semiconductor engineering, novel computer architectures, and a reimagination of deep learning algorithms. I'm excited about all the above!
As much as I love learning about silicon, I actually spend most of my time in software. In college, I worked on light fields, computational photography, and vision-based robotic autonomy, to name a few highlights. I'm very comfortable with building and tuning deep learning models, and have strong familiarity with modern techniques in applied ML.
In my free time, I like to get some use from my EE degree by designing embedded electronics, usually for robotics. I'm also an amateur sysadmin, managing a small-scale server farm designed to get big FLOPs on the cheap.