Mr. Aswathnarayan Muthukrishnan Kirubakaran

Sustainable Data Infrastructure & Edge AI for Emerging Systems

Abstract:
Modern devices such as wearables, AR glasses, and IoT sensors now generate continuous streams of data. Processing all of this data centrally in the cloud introduces challenges related to latency, cost, privacy, and energy consumption. This session focuses on how Edge AI and sustainable data infrastructure can address these challenges by moving computation closer to where data is produced. The talk will walk through a practical reference architecture for building real-time data and machine learning pipelines that operate efficiently at the edge. Key topics include fundamentals of edge-based data processing, strategies for lightweight feature extraction, adaptive sampling to reduce unnecessary data transmission, privacy-preserving model deployment, and telemetry feedback loops for continuous improvement. Attendees will learn how to evaluate trade-offs across latency, resource usage, and model performance, and how sustainability metrics can be incorporated directly into system design decisions. This session is intended for data engineers, system architects, and AI practitioners interested in developing efficient and reliable AI systems for emerging edge-compute environments.  

Profile:

Aswathnarayan is a Data Engineer at Meta Reality Labs. He focuses on building data and AI systems that support wearable and neural interface technologies. His work includes large-scale telemetry processing, real-time analytics, and improving the efficiency and reliability of AI at the device and edge level. Prior to Meta, he developed cloud and data automation platforms in financial and enterprise environments. He is interested in practical, sustainable approaches to AI system design and in helping teams move from experimentation to production deployment.