Crunch Academy · Tech Track
All the mathematics a data scientist or ML engineer needs — taught from the ground up, free.
The Track
Start with Foundations, then take Linear Algebra, Calculus, Probability and Discrete Math in parallel; Optimization and the applied courses build on those. Every lesson is self-contained with worked examples in Python (and R/Julia where it helps).
Rebuild the core math that every data scientist secretly relies on, with Python alongside every concept.
7 units · 48 lessons · 144 worked examples
Open MDS-1 → MDS-2Vectors, matrices, and the geometry that powers machine learning — built up to SVD and PCA.
7 units · 42 lessons · 126 worked examples
Open MDS-2 → MDS-3Limits, derivatives, integrals, series, and numerical methods — the change-and-accumulation toolkit that powers optimization and modeling in data science.
7 units · 44 lessons · 131 worked examples
Open MDS-3 → MDS-4The math behind gradients, Jacobians, and backpropagation — built for machine learning.
7 units · 45 lessons · 135 worked examples
Open MDS-4 → MDS-5Reason about uncertainty the way data scientists do — from coin flips to the Central Limit Theorem, simulated in code.
7 units · 42 lessons · 126 worked examples
Open MDS-5 → MDS-6Turn data into defensible conclusions — estimate, test, and quantify uncertainty with confidence intervals, the bootstrap, A/B tests, regression, and Bayes, all in code.
7 units · 42 lessons · 126 worked examples
Open MDS-6 → MDS-7The math that computers actually run on — logic, counting, graphs, and number theory, every idea grounded in Python code.
7 units · 49 lessons · 119 worked examples
Open MDS-7 → MDS-8Turn the math you know into algorithms a computer can actually run — fast, stable, and accurate enough to trust.
7 units · 48 lessons · 99 worked examples
Open MDS-8 → MDS-9Turn 'find the best answer' into math you can solve — from convex sets and gradient descent to Lagrange multipliers, LPs, and the loss landscapes behind machine learning.
7 units · 45 lessons · 108 worked examples
Open MDS-9 → MDS-10Measure surprise, quantify shared information, and meet the loss functions that train modern models — all built from one number: bits.
5 units · 30 lessons · 90 worked examples
Open MDS-10 →