Leon Zhao

I create physics-informed machine learning models to help solve environmental and natural-hazard problems. I am currently working on an avalanche forecaster that uses satellite data to estimate risk from snowpack, weather, and terrain data at a 2 to 5 km resolution. My model placed 3rd in Engineering at the ISEF-affiliated Golden Gate STEM Fair. I snowboarded for 8 years, and this project started as a way to bridge data science to the mountains I ride.

This past summer I worked as a snow hydrology research intern at UC Santa Barbara's Earth Research Institute, reconstructing snow water storage for California's Cosumnes River watershed. The 2021 Caldor Fire burned its upper basin, and the work supported post-fire water supply research at Lawrence Berkeley National Laboratory. I found a precision bug in the forest canopy classification that had been inflating April 1 snow estimates. That fix and several others I contributed were reviewed and merged into the public codebase.

Outside of snow work, I was in charge of AI and software for LitterBot, an autonomous waste sorting robot that reached the Diamond Challenge Global Finals and won the Delaware Solid Waste Authority (DSWA) Innovation Award. I co-founded my school's AI Club and run a free global summer program teaching Python, Numpy, regression models, and neural networks.