This talk focuses on how AI is practically applied within decentralized infrastructure systems, turning distributed data into real-time decisions across physical environments. It expands how AI models are deployed beyond centralized cloud environments— at the edge (devices, sensors, and local nodes)—to enable low-latency inference, autonomous operation, and efficient resource utilization.
Using real-world case studies in energy, smart cities, and healthcare, the session demonstrates how applied AI enables:
- improve efficiency and resource utilization: real-time prediction and optimization,
- reduce operational costs: automation of infrastructure operations with minimal human intervention,
- increase system resilience and uptime: distributed model training and inference across heterogeneous environments,
- enhance data privacy and regulatory compliance: privacy-preserving AI, where sensitive data remains local.
Overall, the talk provides a practical perspective on building intelligent, resilient, and scalable infrastructure, where AI is embedded directly into the physical world.