This paper describes ML Blocks, https://tinyurl.com/ml-blocks, a novel interface for training, evaluating, and deploying Tiny Machine Learning (TinyML) models. TinyML is a fast-growing field that incorporates powerful machine learning algorithms into everyday technologies such as activity trackers and Internet of Things devices. Although TinyML-capable microcontrollers are popular in computer science education, few students have had the opportunity to learn about the field because of a lack of novice-friendly ML interfaces. With ML Blocks, users assemble data sets, define, and train neural network classifiers, within one unified block interface. Users can quickly evaluate their classifiers using built-in visualization tools and then export them for use in microcontroller projects. ML Blocks makes the end-to-end development of TinyML models easier for physical computing students and tinkerers at all levels.