Bio:

Dr. Mukesh V. Khare is General Manager of IBM Semiconductors division and Vice President of Hybrid Cloud research. He leads a global team of more than 1000 researchers and engineers that are re-defining the future of computing for next generation workloads such as generative artificial intelligence (AI), high-performance computing, and their delivery via hybrid multi cloud. His research agenda includes semiconductor process technology, chip design for accelerated computing, system architecture and hybrid cloud software. Throughout his career, Dr. Khare has built and driven collaborative research alliances that lead to breakthrough advances in computing, such as 7nm and 2nm chip technologies. In 2019, he championed the formation of the AI Hardware Center, a $2.3B initiative to drive innovations in AI hardware technologies through public-private partnerships.

He serves on the Board of the Semiconductor Research Corporation (SRC), which manages more than $100 million in funding for university research in the fields of semiconductors. He is also a member of key academic boards such as Purdue and the State University of New York where he focuses on helping to build the semiconductor industry workforce. In 2021, Mukesh was the recipient of the Asian American Executive of the year award for his vision and execution of breakthrough nanotechnology. Dr. Khare served as the General Chair of the 2018 Symposia on VLSI Technology, has co-authored more than 100 research papers and holds many U.S. and international patents.

Dr. Khare is a recipient of an IBM Corporate Award, IBM’s highest technical award for his leadership in delivering innovative technologies and is also an IBM Distinguished Engineer. He holds a Master of Technology from IIT Bombay and a Doctorate in Electrical Engineering from Yale University. He is a strong advocate of diversity and inclusion in the workplace, and sponsors initiatives such as PowerUp for women engineers.

Abstract:

The potential of generative AI to radically change how we work and live has sparked our collective imagination. These new models can improve productivity through helping us write better code, give us a better customer service experience, or even help to discover new drug therapies.  Just last month at Think 2023, IBM announced watsonx, our full stack approach to providing enterprises with an end-to-end solution for building, tuning, and deploying generative AI solutions.  At the foundation of this stack, IBM is building AI infrastructure optimized for complex AI workloads. In this talk, we will discuss how IBM Research is powering this infrastructure, touching on our Vela supercomputer and our Artificial Intelligence Unit (AIU) based AI System One. We will also cover our roadmap for AI hardware innovation, fueled by what’s next in semiconductor technology.