Machine learning on resource-constrained embedded devices, IoT platforms, and edge computing systems for real-time, low-latency AI deployment.
Edge AI brings artificial intelligence directly to edge devices—smartphones, IoT sensors, microcontrollers, and embedded systems—enabling real-time inference without cloud connectivity. TinyML (Tiny Machine Learning) focuses on ultra-low-power ML models running on microcontrollers with KB-scale memory, perfect for battery-powered IoT applications.
Keywords: edge-ai, tinyml, embedded-ml, on-device-ai, iot-ml, microcontroller-ml, low-power-ai, real-time-inference, embedded-systems, arduino-ml, raspberry-pi-ml, tensorflow-lite, edge-impulse, model-compression
Skill Levels: 🟢 Beginner | 🟡 Intermediate | 🔴 Advanced
- TinyML fundamentals and constraints
- Model optimization for edge devices
- Quantization and model compression
- Embedded hardware platforms (Arduino, Raspberry Pi, ESP32, STM32)
- Edge inference frameworks (TensorFlow Lite, TFLite Micro, Edge Impulse)
- Low-power AI chip architectures
- Real-time sensor data processing
- IoT and edge computing architectures
- Energy-efficient neural networks
- Computer vision on embedded devices
- Federated learning for edge AI
-
Harvard CS249r: TinyML Fundamentals on EdX – World-renowned free Harvard course covering fundamentals of Tiny Machine Learning including applications, algorithms, hardware, and hands-on deployment on microcontrollers. Taught by top experts from Harvard and Google. (🟢 Beginner)
- 📖 Access: Free audit available (certificate optional paid)
- 🏛️ Authority: Harvard University + Google
- ⏱️ Duration: Self-paced
- 🛠️ Hands-on: Yes, with real hardware exercises
- [Tags: beginner tinyml harvard mooc embedded-ml 2025]
- [Verified: 2025-12-16]
-
TinyMLedu Free Online Courses – Comprehensive collection of free TinyML courses and tutorials covering edge AI deployment concepts, programming microcontrollers, and building real-world embedded ML applications with step-by-step guidance. (🟢🟡 Beginner-Intermediate)
- 📖 Access: Fully open, no signup required
- 🌍 Global: Accessible worldwide
- [Tags: beginner intermediate tinyml free-course deployment 2025]
- [Verified: 2025-12-16]
-
Microsoft AI Show: Recognizing Words on Microcontroller using TinyML – Video tutorial demonstrating wake word detection on microcontrollers, training model in cloud using Azure ML Studio, compressing to 18KB, and running on Adafruit EdgeBadge drawing <1mW power. (🟢 Beginner)
- 📖 Access: Fully open (Microsoft Learn)
- 🏛️ Authority: Microsoft AI Show
- 📺 Video: Step-by-step tutorial with demo
- 🛠️ Hands-on: Yes, with EdgeBadge hardware
- ⏱️ Duration: 12 minutes
- [Tags: beginner tinyml wake-word azure-ml microcontroller 2020]
- [Verified: 2025-12-16]
-
Talent500: What is TinyML? Complete Guide to Edge Devices – Comprehensive introduction to TinyML covering machine learning on edge devices, model training/optimization workflow, deployment using TensorFlow Lite, real-world applications, and industry adoption across healthcare, manufacturing, automotive. (🟢 Beginner)
- 📖 Access: Fully open blog article
- 🏛️ Authority: Talent500
- [Tags: beginner tinyml edge-inference iot-applications introduction 2025]
- [Verified: 2025-12-16]
-
JanBask Training: TinyML and Edge AI - Running ML on Small Devices – Article explaining TinyML and Edge AI fundamentals, tools (TensorFlow Lite, PyTorch Mobile, Edge Impulse), real-world projects, required skills, and career opportunities in embedded AI for 2025. (🟢 Beginner)
- 📖 Access: Fully open blog article
- 🏛️ Authority: JanBask Training
- [Tags: beginner edge-ai career-guide tools projects 2025]
- [Verified: 2025-12-16]
-
Edge Impulse: Building ML Models for Embedded Devices (Official Guide) 🟢 Beginner - Official Edge Impulse platform tutorial for building and deploying ML models on microcontrollers without coding. Platform-agnostic guide covering data collection, signal processing, model training, and edge deployment. Includes free tier with hands-on labs using real Arduino, Raspberry Pi, and ESP32 hardware. Perfect for beginners wanting no-code ML deployment.
- 📖 Access: Fully free (official platform)
- 🏛️ Authority: Edge Impulse (official platform provider)
- 🛠️ Hands-on: Yes (free tier with real hardware)
- 🌍 Global: Fully accessible worldwide
- 📊 Tech Stack: No coding required, visual builder, TensorFlow Lite inference
- [Tags: beginner tinyml edge-impulse microcontrollers no-code hands-on 2025]
- [Verified: 2025-01-15]
-
TinyML Applications on Microcontrollers: Revolutionizing Edge AI (ThinkRobotics Guide 2025) 🟢 Beginner - Comprehensive June 2025 guide exploring TinyML applications from voice recognition to predictive maintenance. Covers key technologies (TensorFlow Lite Micro, Edge Impulse, CMSIS-NN, Arduino ML Libraries), popular microcontrollers (ARM Cortex-M series, Arduino Nano 33 BLE Sense, ESP32), and complete getting-started workflow from data collection to model deployment. Includes real-world application examples across wearables, smart home devices, industrial IoT, healthcare monitoring, and environmental sensing.
- 📖 Access: Fully open blog article
- 🏛️ Authority: ThinkRobotics (robotics education platform)
- 📊 Topics: TinyML applications, voice recognition, gesture control, anomaly detection, predictive maintenance, energy harvesting
- 🌍 Global: Fully accessible worldwide
- [Tags: beginner tinyml applications microcontrollers edge-impulse arduino esp32 real-world-examples 2025]
- [Verified: 2026-02-18]
-
AI on the Edge: Deploying TinyML on Microcontrollers (DeepTech Hackathon 2026 Webinar) 🟢 Beginner - Recent February 2026 webinar introducing TinyML fundamentals for the DeepTech Hackathon 2026. Covers what TinyML is and why it matters, popular hardware platforms (Arduino, ESP32, STM32), training and quantizing models using TensorFlow Lite, deploying ML models to microcontrollers, and real-world applications. Ideal entry point for understanding edge AI deployment before diving into hands-on development.
- 📖 Access: Fully free (YouTube webinar)
- 🏛️ Authority: DeepTech Hackathon 2026 (official educational webinar)
- 📺 Video: Educational webinar with slides and demonstrations
- 📊 Topics: TinyML introduction, hardware platforms, model quantization, deployment workflow, applications
- ⏱️ Duration: Webinar format
- [Tags: beginner tinyml introduction webinar hardware deployment applications 2026]
- [Verified: 2026-02-18]
-
TinyML: Getting Started with TensorFlow Lite for Microcontrollers (Digi-Key Electronics YouTube) 🟢 Beginner - Step-by-step video tutorial teaching complete TensorFlow Lite Micro workflow for embedded systems using STM32CubeIDE. Covers training neural networks in TensorFlow/Keras, converting to TensorFlow Lite format, generating C header files for embedded deployment, integrating TFLite Micro library into STM32 projects, and running inference on microcontrollers. Includes practical measurements of flash memory, RAM usage, and inference timing. Perfect hands-on introduction to embedded ML development.
- 📖 Access: Fully free (YouTube + written tutorial)
- 🏛️ Authority: Digi-Key Electronics (official electronics distributor educational content)
- 📺 Video: Complete hands-on tutorial with code walkthrough
- 🛠️ Hands-on: Yes (STM32, Nucleo boards, source code provided)
- 📊 Topics: TensorFlow Lite Micro, model conversion, C integration, STM32CubeIDE, embedded inference, memory optimization
- ⏱️ Duration: Comprehensive video tutorial
- 🌍 Global: Fully accessible worldwide
- [Tags: beginner tinyml tensorflow-lite stm32 microcontrollers hands-on tutorial embedded-inference 2020]
- [Verified: 2026-02-24]
-
TensorFlow Lite (LiteRT): Official Documentation (Google AI) 🟢 Beginner - Official comprehensive documentation for TensorFlow Lite (now rebranded as LiteRT), Google's solution for on-device machine learning. Covers model conversion from TensorFlow to optimized mobile/edge format, quantization techniques for model compression, platform-specific deployment guides for Android/iOS/embedded devices, hardware acceleration with GPU delegates and Edge TPU, and complete API references. Essential resource for understanding Google's edge AI stack with quickstart guides, code examples, and best practices for production deployment.
- 📖 Access: Fully free (official Google documentation)
- 🏛️ Authority: Google AI for Developers (official source)
- 🛠️ Hands-on: Yes (code examples, tutorials, Colab notebooks)
- 📊 Topics: Model conversion, quantization, mobile deployment, hardware acceleration, API reference, optimization
- 🌍 Global: Fully accessible worldwide, multi-platform support
- [Tags: beginner intermediate tensorflow-lite litert google android ios embedded official-docs model-optimization 2026]
- [Verified: 2026-02-24]
-
Computer Vision with TinyML: Edge AI + CNNs (TheTechSlayer, 2025) 🟡 Intermediate - Comprehensive 2025 video course teaching computer vision and TinyML deployment on edge devices. Covers CNN fundamentals, image classification, object detection, and real-world deployment on microcontrollers and Raspberry Pi. Created in collaboration with Edge Impulse, OpenMV, and Seeed Studio. Features practical projects with hands-on hardware labs using industry-standard tools and platforms.
- 📖 Access: Fully free (YouTube)
- 🏛️ Authority: TheTechSlayer⚡💻 (TinyML Foundation collaborators: Edge Impulse, OpenMV, Seeed Studio)
- 📺 Video Series: Complete CV + TinyML course
- 🛠️ Hands-on: Yes (Edge Impulse, OpenMV, Raspberry Pi projects)
- ⏱️ Duration: 6+ hours (complete course)
- 🌍 Global: Fully accessible worldwide
- 📊 Topics: CNN architecture, quantization, edge deployment, computer vision
- [Tags: intermediate tinyml computer-vision cnn edge-impulse openMV raspberry-pi hands-on 2025]
- [Verified: 2025-01-15]
-
ARM Learning Paths: Edge AI with TensorFlow Lite (Official) 🟡 Intermediate - Official ARM learning path for deploying TensorFlow Lite models on microcontrollers and edge devices. Covers model optimization, quantization, inference on ARM Cortex processors, and hardware-specific acceleration. Includes hands-on labs with real ARM-based hardware (Arduino, Raspberry Pi, STM32). Comprehensive resource from processor manufacturer.
- 📖 Access: Fully free (official learning path)
- 🏛️ Authority: ARM (processor manufacturer, official source)
- 🛠️ Hands-on: Yes (interactive labs with real hardware)
- 📊 Topics: Model quantization, TensorFlow Lite Micro, ARM Cortex optimization
- 🌍 Global: Fully accessible worldwide
- ⏱️ Duration: Self-paced learning path
- [Tags: intermediate tinyml tensorflow-lite quantization arm-cortex hands-on 2025]
- [Verified: 2025-01-15]
-
MathWorks: Rapid Deep Learning Deployment on Edge Devices – Technical guide for streamlining TinyML application development on resource-constrained devices using MATLAB, covering network compression, quantization, pruning, and benchmarking on STMicroelectronics edge hardware. (🟡 Intermediate)
- 📖 Access: Fully open technical article
- 🏛️ Authority: MathWorks + STMicroelectronics
- 🛠️ Hands-on: MATLAB-based workflow
- [Tags: intermediate edge-deployment model-compression benchmarking 2025]
- [Verified: 2025-12-16]
-
Edge AI Foundation Talks & Livestreams – Technical talks and discussions on on-device tiny machine learning, addressing memory, computation, and energy efficiency constraints with adaptive mechanisms for real-time inference and on-device training strategies. (🟡🔴 Intermediate-Advanced)
- 📖 Access: Free livestreams and recordings
- 🏛️ Authority: Edge AI Foundation
- [Tags: intermediate advanced tinyml on-device-learning technical-talks 2025]
- [Verified: 2025-12-16]
-
Edge Impulse: tinyML AutoML Deep Dive Tutorial (YouTube) – 44-minute hands-on video tutorial demonstrating Edge Impulse's EON Tuner for automated model selection, analyzing input data, signal processing blocks, and neural network architectures within device latency and memory constraints. (🟡 Intermediate)
- 📖 Access: Fully open (YouTube)
- 🏛️ Authority: Edge AI Foundation (official)
- 📺 Video: Complete tutorial with live demo
- 🛠️ Hands-on: Yes, Edge Impulse Studio
- ⏱️ Duration: 44 minutes
- [Tags: intermediate automl edge-impulse tinyml tutorial hands-on 2022]
- [Verified: 2025-12-16]
-
Noble Prog: Federated Learning for Privacy-Preserving AI Training – Training course covering federated learning implementation for privacy-protected AI models on edge devices, including differential privacy, secure aggregation, encryption techniques, and real-world healthcare applications. (🟡 Intermediate)
- 📖 Access: Course outline fully open
- 🏛️ Authority: Noble Prog Training
- 🛠️ Hands-on: Case studies and exercises
- [Tags: intermediate federated-learning privacy edge-ai healthcare 2025]
- [Verified: 2025-12-16]
-
Deploying AI on Microcontrollers with TinyML (NobleProg Training Course) 🟡 Intermediate - Comprehensive instructor-led training course (online or onsite) for embedded systems engineers and AI developers teaching practical TinyML deployment on microcontrollers using TensorFlow Lite and Edge Impulse. Covers TinyML fundamentals, development environment setup for Arduino IDE and TensorFlow Lite, training and optimizing AI models for low-power devices, real-world deployment on microcontrollers, and tuning for power efficiency and memory constraints. Hands-on approach with sensor data collection, model training, and edge deployment exercises.
- 📖 Access: Free course outline fully accessible (training sessions available)
- 🏛️ Authority: NobleProg (established technical training provider)
- 🛠️ Hands-on: Yes (practical labs, microcontroller deployment)
- 📊 Topics: TinyML workflow, TensorFlow Lite Micro, Edge Impulse Studio, model optimization, power efficiency, Arduino IDE
- ⏱️ Duration: Instructor-led format (flexible scheduling)
- [Tags: intermediate tinyml tensorflow-lite edge-impulse microcontrollers hands-on training 2025]
- [Verified: 2026-02-18]
-
Deploying Machine Learning on Microcontrollers: Complete TinyML Guide (SmartTech Fusion YouTube 2025) 🟡 Intermediate - Comprehensive video guide covering the complete TinyML deployment pipeline from training workstations to resource-constrained microcontrollers. Explores model quantization techniques for memory/compute optimization, deployment pipelines using TensorFlow Lite Micro and Edge Impulse, real-time inference on Arduino/ESP32/STM32 platforms, and applications in smart agriculture, predictive maintenance, and gesture recognition. Includes performance optimization techniques for low-power operation and practical considerations for production deployment.
- 📖 Access: Fully free (YouTube video tutorial)
- 🏛️ Authority: SmartTech Fusion (embedded systems education channel)
- 📺 Video: Complete deployment guide with demonstrations
- 🛠️ Hands-on: Yes (Arduino, ESP32, practical examples)
- 📊 Topics: Model quantization, TFLite deployment, microcontroller platforms, optimization, real-world applications
- ⏱️ Duration: Comprehensive tutorial
- 🌍 Global: Fully accessible worldwide
- [Tags: intermediate tinyml deployment quantization arduino esp32 tensorflow-lite applications optimization 2025]
- [Verified: 2026-02-24]
-
TinyML Tutorials: Enabling Intelligent Edge Devices (GitHub) 🟡 Intermediate - Open-source GitHub repository with hands-on tutorials for deploying machine learning on ultra-low-power ARM Cortex-M microcontrollers using TensorFlow Lite. Includes complete walkthrough of training CNNs for CIFAR-10 image classification, converting models to TensorFlow Lite format, integrating TFLite Micro with Mbed development environment, and deploying to STM32H747I Discovery Kit. Features code examples, build instructions, and performance benchmarking. Covers CMSIS-NN optimizations for ARM SIMD instructions (Cortex-M4/M7/M33/M35P) to achieve significant inference speedups.
- 📖 Access: Fully free (open-source GitHub repository)
- 🏛️ Authority: Community-maintained (ARM ecosystem)
- 🛠️ Hands-on: Yes (complete code repository, STM32 hardware examples)
- 📊 Topics: ARM Cortex-M, TensorFlow Lite Micro, CIFAR-10, CNN deployment, CMSIS-NN, Mbed, STM32
- 💻 Tech Stack: C++, TensorFlow Lite Micro, Mbed CLI, ARM cross-compilation toolchain
- 🌍 Global: Fully accessible worldwide
- [Tags: intermediate tinyml arm-cortex tensorflow-lite stm32 hands-on github cifar-10 cmsis-nn 2020]
- [Verified: 2026-02-24]
- Secure and Private AI: Differential Privacy and Federated Learning (PDF) – Academic presentation slides covering differential privacy fundamentals, federated learning architecture, privacy guarantees, and mathematical frameworks for secure edge AI systems with minimal accuracy loss. (🔴 Advanced)
- 📖 Access: Fully open (PDF slides)
- 🏛️ Authority: Washington State University
- 📄 Format: Technical presentation (87 slides)
- [Tags: advanced differential-privacy federated-learning mathematical research 2019]
- [Verified: 2025-12-16]
Popular Edge AI Frameworks:
- TensorFlow Lite (LiteRT) - Lightweight ML for mobile and embedded devices (Google)
- TensorFlow Lite Micro - Ultra-lightweight for microcontrollers
- Edge Impulse - End-to-end platform for embedded ML (free tier available)
- ONNX Runtime - Cross-platform inference optimization
- Arduino TensorFlow Lite - TinyML for Arduino boards
- STM32Cube.AI - Neural network optimization for STM32 MCUs
- PyTorch Mobile - Mobile deployment for PyTorch models
- Apache TVM - Deep learning compiler for edge devices
- OpenMV - Computer vision on embedded systems
- CMSIS-NN - ARM neural network kernels for Cortex-M processors
Hardware Platforms:
- Arduino Nano 33 BLE Sense
- Raspberry Pi (all models)
- ESP32 / ESP32-S3
- NVIDIA Jetson (Nano, Xavier)
- Google Coral Edge TPU
- STM32 microcontrollers (STM32H747I Discovery Kit)
- Adafruit EdgeBadge
- Seeed Studio boards
- OpenMV camera boards
- ARM Cortex-M series (M4, M7, M33, M35P)
See also:
- Deep Learning & Neural Networks - Models to deploy on edge
- Computer Vision - Vision models for edge cameras
- MLOps - Deployment and monitoring practices
- AI Tools & Frameworks - ML frameworks and libraries
- AI Security & Privacy - Federated learning and differential privacy
Cross-reference:
- Graph Neural Networks - Edge AI in IoT networks
- Audio & Speech Processing - Speech recognition on edge devices
- Reinforcement Learning - On-device learning and adaptation
Found a great free Edge AI or TinyML resource? We'd love to add it!
To contribute, use this format:
- [Resource Name](URL) – Clear description highlighting value and what you'll learn. (Difficulty Level)
- 📖 Access: [access details]
- 🏛️ Authority: [Source/Organization]
- [Tags: keyword1 keyword2 keyword3 year]
- [Verified: YYYY-MM-DD]
Ensure all resources are:
- ✅ Completely free to access (no payment required)
- ✅ Openly available (no authentication barriers for core content)
- ✅ High-quality and educational
- ✅ Relevant to Edge AI / TinyML
- ✅ From reputable sources (official docs, universities, established platforms)
- ✅ HTTP 200 verified before submission
Last Updated: February 24, 2026 | Total Resources: 21 (+4 new) Last Link Validation: February 24, 2026
Keywords: edge-ai, tinyml, embedded-machine-learning, on-device-ai, iot-ml, microcontroller-ml, low-power-ai, tensorflow-lite, litert, edge-impulse, arduino-ml, real-time-inference, model-compression, quantization, embedded-systems, raspberry-pi, federated-learning, privacy-preserving-edge-ai, computer-vision-edge, openMV, arm-cortex, stm32, esp32, cmsis-nn, digi-key, google-ai, github-tutorials, deployment-pipeline, 2025-2026, free-courses