The use of ML in embedded systems allows devices to process and immediately act upon data in a much different way from before. When embedded into microcontrollers or edge appliances, computer vision, anomaly detection, and prognostics processes do not require cloud support. It is common for projects here to attempt to use models that are low power and have small memory requirements so they are well-suited for IoT, Robotics, and Smart Devices. From gesture recognition to voice control to high efficient data handling and everything in between, Machine Learning in embedded systems brings new concepts with a performance-driven and real-time oriented perspective.