Hyunwon Chung

Hello! My name is Hyunwon Chung, and I am a PhD student at the University of Michigan. I am supervised by Professors David Blaauw and Hun-Seok Kim.

My research interests span reconfigurable and dataflow architecture for wireless communication and machine learning. I have taped out multiple chips across GF 12nm and TSMC 28nm processes — including heterogeneous SoCs for beyond-5G optical communication and streaming processors for advanced spectrum sensing.

More recently, my work has focused on efficient hardware for large language models. I design hardware architectures tailored for approximation techniques in billion-parameter LLMs, develop performance and energy models to analyze the hardware impact of approximation algorithms, and perform architecture-level design space exploration to identify efficient accelerator configurations for approximated LLM workloads.