About

I'm Shao — an engineer who likes the layer where software has to answer to physics. I studied electrical and computer engineering at Cornell (Class of 2026), where my favorite work involved making real hardware do difficult things: a differential-drive robot that Kalman-filters its way past its own sensor latency, the autonomous side of a 3 lb combat robot — I led the autonomy subteam at Cornell Combat Robotics — and a sparse-matrix kernel pushed to ~21× speedup on 32 cores.

From Fall 2024 through Spring 2026 I was an undergraduate researcher in Cornell's EmPRISE Lab, working on vision-language models and diffusion models for robot learning. Today I write embedded software for medical devices at Smith & Nephew (IEC 62304, MicroBlaze soft-core FPGAs) and run a personal training-to-edge ML setup — an RTX 5090 workstation feeding a Jetson Orin Nano — that carries in-progress projects like SpeechLens, a fully local speech-analysis pipeline.

This site is one of those projects: static, fast, and built in the open — the repo is part of the portfolio.