Abstract: In this talk, we will cover the full stack approach from IBM toward Foundation Models, from building our own AI chip, being part of the open AI software ecosystem (e.g., PyTorch and vLLM) to bringing Generative AI capabilities to Enterprises through watsonx. We will discuss the Research challenges across the various parts of the stack and what it takes for all these pieces to work in unison.
Bio (Ganti): Raghu Ganti is a Principal Research Scientist at IBM Research and received his PhD in Computer Science from UIUC (2010). He leads a team at IBM working on PyTorch enhancements and enabling model customization for enterprise workloads in Hybrid Cloud. His past work includes developing a spatiotemporal analytics library that is part of 15+ IBM products.
Bio (Srivatsa): Mudhakar Srivatsa is a distinguished research staff member at the Distributed Cloud department in IBM T. J. Watson Research Center. His work is focused on heterogeneous spatiotemporal data with applications to edge computing, AIOps and Hybrid AI Scaling. He is an IBM master inventor, authored over 200 research papers, 100 granted US patents, recipient of one IBM corporate award, one IBM Recognition Experience Honoree, nine IBM outstanding technical achievement awards and three IBM research division awards and has transitioned major software artifacts to various IBM products. His machine learning algorithms have been used in production environment in various domains such as: CodeFlare (blogs, github), AIOps, remote sensing data, AI-assisted air traffic control (video), NASA space app challenge (github, video, blog), customer care analytics, maritime piracy, music festivals, connected vehicles, smart wildlife, and predicting asteroid encounters.
Bio (Srinivasan): Viji Srinivasan is a Research Staff at IBM T.J. Watson Research Center in Yorktown Heights. She joined IBM in 2001 after completing her PhD at the University of Michigan. Her research interests are in computer architecture, with particular focus on processor microarchitecture, and multi-core/multiprocessor memory systems. She has received an IBM Outstanding Technical Achievement Award for her work on power-aware microarchitectures, an IBM Technical Group Award for her contribution to the PERCS architecture.
