Abstract

Verifiable evidence is the foundation of reliable AI-assisted root cause analysis (RCA). This talk combines AI agents with time-travel debugging to investigate failures in complex EDA software. Agents trace failures through recorded execution and source code; engineers validate intended behavior and make sure the delivered fixes are strategic, rather than tactical. Each reviewed investigation can turn architectural knowledge into explicit behavioral contracts and regression tests, building a foundation for safe automation. The talk presents a staged path toward autonomous bug fixing, with autonomy gated by specification clarity, execution evidence, and reliable checks.

Bio

Chau-Chin Huang received his B.S. degree in electrical engineering from National Cheng Kung University (NCKU), Tainan, Taiwan, in 2012, and his Ph.D. degree in electrical engineering from National Taiwan University (NTU), in 2017. He is currently a Research and Development Engineer with Design Group, Synopsys, Hsinchu, Taiwan. His research interests include Agentic Debugging, VLSI physical design, and placement.

All faculty and students are welcome to join.