Yuzhe Ma of The Hong Kong University of Science and Technology (Guangzhou)

Photonic integrated circuits (PICs) are emerging as a key technology for high-bandwidth communication, energy-efficient interconnects, and next-generation computing, yet their design still relies heavily on expert knowledge and labor-intensive iterations. This talk presents recent advances toward PIC design automation across different layers, spanning device-level inverse design, circuit-level placement and routing, and system-level modeling. It also discusses how mathematical optimization, AI, and large language models can complement conventional algorithms to improve design efficiency while respecting the distinctive constraints of photonics. Together, these efforts point toward scalable, reliable, and manufacturable design automation across abstraction levels, while highlighting open challenges on the path to truly end-to-end photonic design.