AI on the Edge – Unlocking New Opportunities in IoT
summary
Demonstrates running containerized models with Azure IoT Edge to orchestrate and securely deploy AI from Raspberry Pi–class boards up to GPU servers, using TPM/x.509 provisioning and Azure Stream Analytics for on‑device inference. Live demo runs YOLO on Jetson Nano, detects cup/person and flags a mis-sorted package, proving a ~\$100 device can process one device handling multiple 1080p streams and send only selected events to the cloud, with GDPR face‑blurring and offline, low‑latency operation.
AI-written summary · about
chapters (14)
- 0:00Start
- 2:14What do these scenarios have in common?
- 4:11Edge is hard
- 6:19lot application pattern + edge intelligence
- 6:58Edge intelligence enabled with Azure lot Edge
- 8:42Basic Azure lol Edge functionality
- 10:02Enabling intelligent edge spectrum
- 11:15Azure loT Edge features
- 12:32Telemetry
- 14:15Built-in Anomaly Detection with Azure Stream Analytics at the Edge
- 18:51Examples of Vision-capable Edge Devices
- 22:433 Paths to get an Al Model
- 24:002 architectures each with their trade-offs
- 29:59Manage your data
related talks
01h 00mDrones, AI & IOT - What's all the buzz about ?
24mWhat does AI mean for IoT development?
27mGenerative AI and IoT, is there a match?
18mThe Cutting Edge of AI and IoT
14mEdge AI: Real Solutions for Real World Problems
05mEdge AI: Because the Cloud is Too Slow...
50mAzure AI, Power new possibilities for every organization
15mSimplifying & Scaling Edge AI