AI By Doing is an educational brand created by huhuhang, offering a practical, continuously updated tutorial series for learning machine learning and artificial intelligence. Its central resource combines clear explanations of algorithmic principles with mathematical derivations and Python implementations, helping learners connect theory with working code. Lessons also demonstrate how to apply mainstream open-source frameworks, including scikit-learn, and compare library-based solutions with implementations built from the fundamentals.

The curriculum covers supervised learning, including regression and classification, as well as unsupervised learning topics such as clustering and principal component analysis. Practical projects explore applications including housing-price prediction, image classification, image compression, and analysis of shared-bike GPS data. Materials are presented in a format suited to hands-on practice with Jupyter Notebook, and supplementary references cover mathematical foundations and commonly used tools. The site currently contains 97 chapters and is still being updated. All tutorial content is available to read for free. Readers may optionally support its maintenance through a Pro membership; the creator also offers independently developed iOS apps.