Talk title

Driving Pervasive Intelligence with AI-Driven EDA.

Presentation

AI in chip design is evolving from a “Question to Answer” paradigm where engineers pose questions and AI provides responses to a “Goal to Result” approach, where engineers define objectives and autonomous agents deliver outcomes. This talk explores the progression from reinforcement learning to generative copilots and, ultimately, to fully orchestrated multi-agent systems. It highlights how purpose-built agentic architectures achieve 35% higher accuracy, 40% lower token costs, and 60% lower latency compared to traditional LLM-based methods. We showcase how autonomous agents effectively plan, reason, and complete design tasks
across RTL, verification, and signoff, and discuss the strategic implications for engineering teams.

Biography

Asi Sapir is a hardware engineer and AI specialist at Synopsys, where he serves as a Power End-to-End and AI Specialist driving the adoption of AI-powered EDA solutions. With 25 years of experience at Intel and 3 years at Synopsys, he brings deep expertise in logic design, RTL methodology, architecture, physical design, and design automation. Asi works at the intersection of chip design methodology and artificial intelligence, helping engineering teams integrate reinforcement learning, generative AI, and agentic workflows into production environments. He holds a B.Sc. in Electrical Engineering from the Technion — Israel Institute of Technology.