Member of Technical Staff, eBay Inc, San Jose, California, USA

Title of the Talk :
From Centralized AI to Edge Intelligence: Building Distributed, Context-Aware Systems for the Real World

Abstract of Talk:
As artificial intelligence becomes deeply embedded in everyday experiences—from smart assistants to autonomous vehicles—the traditional model of centralized AI processing is being re-evaluated. Latency, bandwidth, privacy, and resilience concerns are driving a shift toward edge intelligence, where computation occurs closer to the data source. This keynote explores the transition from cloud-centric AI to distributed, context-aware systems that operate on edge devices, mobile networks, and local environments. We will examine the convergence of AI model design, edge computing, and communication protocols, with a focus on how real-time inference, sensor fusion, and adaptive learning are transforming the architecture of intelligent systems. Key technical challenges will be addressed, including model compression, federated and split learning, on-device optimization, and intelligent data routing. The talk also highlights how edge intelligence can enhance privacy, reduce infrastructure costs, and enable new use cases in healthcare, transportation, industrial IoT, and more. By bridging insights from computational intelligence and next-generation data communication, this session presents a blueprint for building responsive, efficient, and ethically grounded AI systems in the real world—systems that do not simply replicate intelligence, but embed it into context.