Aarav Shah

I'm an undergraduate student studying Computer Science at Berkeley. My interests, broadly defined, are at the intersection of deep learning, optimization models, and hardware. Currently, I am learning about GPU architectures and leveraging my knowledge to improve transformer inference at AMD.

I began as a Chemical Engineering major, and had the chance to explore research at the intersection of computation, chemistry, and biology. A few summers ago, I built models to predict the properties of novel ionic liquids utilized in carbon separation. I later developed an ultrasensitive spectrometer to study biomolecular dynamics at the picosecond scale, and most recently, built computational tools to support oral microbiome research.

On the industry side, I've worked across both hardware and biotech. At Tesla, I was on the Autopilot Hardware team optimizing transformer inference for the upcoming Ai5 chip. At Merck, I helped design and train Direct Message Passing Neural Networks (DMPNNs) for small-molecule drug discovery.

I've had the pleasure of growing up first in Ahmedabad, and then Austin. I love trying out recipes, long walks, and reading, as well. At the moment, I've been practicing the electric guitar (a hearty thanks to Arun) and reading books on history & macroeconomics.

Feel free to reach out: aarav at berkeley dot edu