Introduction

The convergence of artificial intelligence and embedded systems is revolutionizing how we interact with technology. TinyML—Machine Learning on microcontrollers—brings intelligent decision-making directly to edge devices, enabling applications from smart homes to industrial automation, all while maintaining privacy, reducing latency, and minimizing power consumption.

This book demystifies TinyML through hands-on projects using Seeed Studio’s powerful yet affordable development platforms: the XIAOML Kit based on the XIAO ESP32S3 Sense and the Grove Vision AI V2. Whether you’re classifying images, detecting objects, recognizing keywords, or identifying motion patterns, you’ll learn by building real, working applications.

The XIAO ESP32S3 Sense, despite its thumb-sized form factor (21 x 17.5mm), packs remarkable capabilities: a dual-core ESP32-S3 processor running at 240 MHz, 8MB of PSRAM, an OV2640 camera, and a digital microphone. Combined with the XIAOML Kit’s expansion board, which features an OLED display and an IMU sensor, this platform provides everything needed for comprehensive TinyML exploration.

The Grove Vision AI V2 offers a complementary approach: a dedicated AI vision module with built-in hardware acceleration, ideal for no-code and low-code computer vision applications. Together, these platforms demonstrate the spectrum of TinyML deployment strategies, from fully custom models to pre-trained solutions.

Throughout this book, you’ll use Edge Impulse Studio as your primary development platform, gaining practical experience in the complete machine learning workflow: data collection, preprocessing, model training, optimization, and deployment. Each project systematically builds your understanding, from foundational concepts to advanced techniques such as digital signal processing and feature engineering.

This isn’t just a technical manual—it’s your practical guide to building intelligent edge devices that can see, hear, and understand the world around them.