As AI workloads increasingly move from the cloud to the edge, organizations face critical challenges in orchestrating intelligent systems across millions of heterogeneous devices. This session explores the architectural patterns and practical strategies for building production-grade Edge AI platforms that operate autonomously while maintaining coordination with cloud-based orchestrators.

Drawing from real-world experience deploying AI systems across industrial IoT, healthcare, and smart infrastructure environments, this talk covers:

Edge AI Architecture Patterns: Designing resilient, distributed systems that leverage CPU/GPU/TPU/NPU heterogeneity

Agentic AI at the Edge: Building multi-agent systems that make autonomous decisions with minimal latency

Orchestration Strategies: Balancing edge autonomy with cloud coordination using modern messaging systems (gRPC, MQTT) and cloud-native technologies (Kubernetes, Docker)

Production Lessons: Key insights from scaling Edge AI platforms to millions of devices.