Projects
100 Days of Math for AI
The mathematics of machine learning, implemented from scratch
- Year
- 2026
- Type
- AI / ML
- Role
- Learning project
- Status
- In development
A public, 100-day study of the mathematics behind machine learning: linear algebra, multivariate calculus, probability and optimisation. The rule is no ML libraries; every idea is implemented in Python and NumPy in Jupyter notebooks.
Approach
The plan runs in four pillars: linear algebra (days 1–30), multivariate calculus (31–55), probability and statistics (56–85), and optimisation with core ML mathematics (86–100). Each day is a notebook that derives an idea and implements it by hand: vectors and norms, matrix products as forward passes, inverses, determinants, rank, null and column spaces.
Technology
- Python
- NumPy
- Jupyter