Speaker:
Vijay Janapa Reddi
Gordon McKay Professor
Harvard University
Talk Title:
Architecture 2.0: Foundations for AI Agents in Computer System Design
Date and Location:
Friday, October 2, 2026
3–4 p.m.
KP 108
There is no registration required to attend this event in person. However, seating is limited, so arriving early is recommended.
Abstract:
Modern computing systems have reached unprecedented levels of complexity, pushing traditional design methodologies to their limits. As architectures become increasingly specialized and heterogeneous, designers must navigate vast, interconnected design spaces spanning hardware, software, compilers, runtimes, and machine learning workloads. This talk presents Architecture 2.0, a vision for the next generation of computer system design in which AI agents become active collaborators throughout the research, optimization, and design process. While machine learning has already influenced architecture through automated optimization and design space exploration, autonomous AI agents have the potential to fundamentally reshape how we explore architectural tradeoffs, optimize hardware and software co-design, and discover entirely new system designs. Realizing this vision requires more than advances in AI. It requires a shared ecosystem of standardized environments, datasets, benchmarks, interfaces, and evaluation methodologies that enable agents to learn, collaborate, and improve. This talk explores the foundational infrastructure needed to make AI agents first class participants in computer architecture research and outlines a community driven path toward building the next generation of intelligent system design.
Biography:
Prof. Vijay Janapa Reddi is a Gordon McKay Professor of Engineering and Applied Sciences at Harvard University, where his research focuses on Physical AI: building the systems, architectures, and measurement infrastructure for AI that operates in the real world. As AI moves into physical environments, his work addresses the critical challenges of safety, efficiency, and rigorous evaluation. He is currently on sabbatical at ETH Zurich, hosted by the Department of Information Technology and Electrical Engineering (D-ITET), where he co-advises Master's theses and semester projects in machine learning systems and AI for system design.
Dr. Prof. Janapa Reddi is widely recognized for his pioneering contributions to the field, including developing the emerging field of Tiny Machine Learning (TinyML) and co-founding MLPerf as a founding member and co-chair of the Inference Working Group—now the industry-standard benchmarking suite that evaluates machine learning systems from megawatt to microwatt scales. These foundational works have shaped how the industry approaches efficient AI deployment and performance measurement.
In addition to his academic role, Dr. Prof. Janapa Reddi is deeply involved in shaping the future of machine learning and edge AI technologies. He serves as Vice President and co-founder of MLCommons, a nonprofit organization dedicated to accelerating machine learning innovation. He also serves on the boards of directors for the EDGE AI Foundation, fostering academic-industry partnerships at the edge of AI.
Dr. Prof. Janapa Reddi is passionate about expanding access to applied machine learning and promoting diversity in STEM. His open-source book "Machine Learning Systems" is widely adopted by institutions worldwide, and his Tiny Machine Learning (TinyML) series on edX has trained over 100,000 students globally, democratizing access to cutting-edge AI education.