Bio:

Guillaume Granet is a Senior Account Engineer with more than 12 years of experience in the RF and test & measurement industry. After seven years in R&D at Safran as an RF Engineer, he joined Keysight Technologies, where he spent five years supporting customers as a Support and Application Engineer. He now leverages strong technical expertise and hands‑on experience to help key accounts address complex measurement challenges and optimize their solutions.

Abstract:

The rapid acceleration of AI adoption is redefining infrastructure requirements across compute, memory, and networking. As AI models increase in scale and complexity, performance and reliability depend on holistic system validation. This inflection point extends beyond raw bandwidth: training and inference workloads are increasingly constrained by latency, synchronization, concurrent access to shared resources, and large scale data movement. In parallel, higher density silicon, advanced packaging, and heterogeneous integration introduce new design trade-offs and failure modes.
This talk explores key challenges shaping modern AI infrastructure, including scaling compute and memory architectures, sustaining predictable network performance, and managing multiple points of failure across complex systems. It examines the full life-cycle from pre-silicon architecture and modeling to packaging and post-silicon validation highlighting where bottlenecks emerge and how the industry is evolving to support AI systems.