Jeff Dyck - Mentor, A Siemens Business, Saskatoon, Saskatchewan, Canada
Speaker Bio: Jeff Dyck is a Director of Engineering at Mentor, A Siemens Business, where he runs the Solido R&D team that was acquired by Mentor in December, 2017. Previously at Solido, Jeff was VP of Technical Operations and was part of the executive team that built Solido from the beginning into one of the fastest growing tech companies in North America. Jeff led Solido’s R&D team, managed Solido’s product lines, and co-invented many of Solido’s machine learning and reinforcement learning technologies, including High-Sigma Monte Carlo, PVTMC Verifier, Hierarchical Monte Carlo, and Statistical Characterizer. These technologies form the foundation of Variation Designer and ML Characterization Suite, which are mainstream chip design tools that are used by thousands of circuit designers across most of the top-40 semiconductor companies.
The Golden Age of machine learning is upon EDA. Over the past four years, we have seen large EDA suppliers and customers grow their internal ML teams and strategies, and ML research projects are emerging in all areas of EDA. But, we have not yet seen much of this investment convert into real production flows and work. This is because it is hard to turn research prototypes into production-ready ML tools that function correctly in the real world. This talk reviews a set of second order challenges that make it difficult to bring ML solutions to production, and discusses approaches for solving them.