ML-For-Beginners (Microsoft)
Free Microsoft curriculum, MIT-licensed, to learn classical machine learning (the kind that predates large language models): 12 weeks, 26 lessons, 52 quizzes. Nine themes, from the introduction to regression, classification, clustering, language processing, time series, reinforcement learning and real-world cases. Each lesson follows the same rhythm: pre-quiz, lesson, knowledge checks, challenge, assignment, post-quiz. Code is mostly Python with Scikit-learn, with R versions for many lessons. Spotted in an [Instagram reel](https://www.instagram.com/reel/DfH5IpFgMln/) by Shirin Jam listing five repos to learn machine learning.
Strengths
- Same rhythm in every lesson, with quizzes before and after to check what stuck
- Free, MIT-licensed, usable in the browser with GitHub Codespaces without installing anything
- Code solutions to compare with your own attempt, the repo asks you to build before looking
Limitations
- Classical machine learning only: nothing on language models, agents or generative AI
- Needs a little Python to run the notebooks, even though the repo sets no formal prerequisite
- Long: 12 weeks at a steady pace, poorly suited to someone who just wants the vocabulary
Best for
- Understanding what regression, classification and clustering mean when a data team cites them
- Getting started with machine learning with a structured frame rather than a list of links