Edge AI Deployment: Running LLMs and Neural Networks on Embedded Systems and IoT Devices - Tapa blanda

Sinopsis

Deploy AI at the Edge — Where Latency Is Zero and Bandwidth Is Zero

The edge AI revolution has arrived. In 2026, quantized language models run on Raspberry Pi 5, neural networks execute on microcontrollers with 256KB of RAM, and industrial IoT gateways perform real-time inference without cloud connectivity. Edge AI Deployment is the definitive engineer's guide to making it work reliably in production.

This comprehensive book covers the full stack from model quantization theory to production firmware. Learn how to shrink a 7-billion-parameter LLM to run on a Jetson Orin Nano, deploy TensorFlow Lite models on ESP32, and benchmark power consumption against accuracy tradeoffs. You will also discover how to manage model updates over constrained networks and secure your deployments.

Inside this guide, you will master:

• Quantization Techniques: INT8, INT4, GPTQ, and AWQ for LLMs to maximize efficiency.
• ONNX Runtime & TensorFlow Lite: Cross-platform inference on ARM Cortex-A/Cortex-M and mobile SoCs.
• Live Hardware Benchmarks: Real-world testing on Raspberry Pi 5, Jetson Orin Nano, and ESP32-S3.
• Power Budgeting & System Design: Balancing battery life, latency, and inference frequency.
• OTA Model Updates: Lifecycle management over MQTT, CoAP, and LTE-M networks.
• Edge Security: Model encryption, secure boot, and anti-tampering for remote devices.

Whether you are retrofitting a legacy embedded system with intelligent capabilities or designing a new ultra-low-power IoT product from the ground up, this book gives you the engineering judgment to deploy AI at the edge with total confidence. Start building the next generation of embedded intelligence today.

"Sinopsis" puede pertenecer a otra edición de este libro.