Eight notebooks that shore up the math ML papers assume you already know

What it does This repo holds eight Jupyter notebooks that make up Jon Krohn’s Machine Learning Foundations curriculum. They cover linear algebra, calculus, probability, statistics, algorithms, data structures, and optimization, all aimed at giving you a practical, functional understanding of the math and CS that sit underneath modern machine learning frameworks. The examples are in Python and focus on NumPy, PyTorch, and TensorFlow.

The interesting bit Krohn structures the material around the “Machine Learning House”: foundational subjects form the base, ML theory is the central purple floor, and specialized domains like NLP or computer vision sit on the gray top floor. The repo is explicitly aimed at people who already use high-level libraries like scikit-learn or Keras but need to read papers or debug production systems without getting lost in the notation.

Key highlights

• Eight sequential subjects across four independent areas: Linear Algebra, Calculus, Probability & Statistics, and Computer Science.
• Heavy use of hands-on Python code alongside paper-and-pencil comprehension exercises with fully-worked solutions.
• Intended to run in the free Colab environment; YouTube lectures cover the full taught content.
• Paid versions on O’Reilly, Udemy, and ODSC include exclusive solution walkthroughs, cheat sheets, and completion certificates.
• A Pearson book is planned but not yet written.

Caveats

• The README incorrectly links the notebooks directory to a different repository (jonkrohn/DLTFpT), suggesting a copy-paste error.
• The Dockerfile’s library versions are explicitly flagged as potentially outdated.
• Local Jupyter and Docker setups are supported only passively; Colab is the only actively maintained environment.

Verdict Worth your time if you use high-level ML libraries and want to understand the fundamentals underlying the abstractions. Less useful if you already possess graduate-level fluency in these subjects or need a quick-reference cheat sheet instead of a structured curriculum.