What is Edge AI?
Edge AI means running artificial intelligence algorithms locally on a device — at the "edge" of the network — rather than sending data to a remote server or the cloud.
Cloud AI vs. Edge AI
| Cloud AI | Edge AI | |
|---|---|---|
| Where it runs | Remote server | On the device/near the device itself |
| Latency | 100 ms – several seconds | < 10 ms (real-time) |
| Internet | Required | Not required |
| Privacy | Data leaves the device | Data stays on device |
| Power | High (server) | Ultra-low (mW range) |
| Cost | Per-API-call billing | One-time hardware |
Why Edge AI Matters
Real-Time Response
A keyword-spotting model on a microcontroller responds in milliseconds. Sending audio to a cloud API and waiting for a response takes seconds — unacceptable for many applications.
Privacy by Design
Medical wearables, security cameras, and industrial sensors often can't (or shouldn't) send raw data over the internet. Edge AI keeps sensitive data local.
Works Offline
Edge devices function in remote locations, underground, or in poor-connectivity environments — no internet required.
Energy Efficiency
A TinyML model on a microcontroller can run for months on a coin cell battery. Cloud-dependent systems require continuous data transmission.
Where is Edge AI Used?
- Industrial predictive maintenance — vibration anomaly detection on factory floors
- Healthcare wearables — ECG classification, fall detection
- Agriculture — soil sensor analysis, pest detection with cameras
- Consumer electronics — face unlock, gesture control, noise cancellation
- Robotics — real-time object detection and navigation
Next Steps
→ Workshop Overview — see the full agenda
→ Arduino Q Setup — get your hardware ready