Foundations

Foundations is where I keep working notes and small, solved problems in “traditional” machine learning and deep learning. These aren’t polished tutorials. They’re the minimal notebooks I want to be able to reopen in a year and still understand end‑to‑end.

Fundamentals

My Super Simple ML Workbench (That Covers ~80% of Classic ML)

For most tabular ML work — loading CSVs, training scikit-learn baselines, plotting results — you need surprisingly little tooling. Here's the boring, reproducible workbench I actually use: uv for environments and deps, VS Code for notebooks and scripts, and nothing else.

Evaluating ML Models: It’s About Choosing Your Mistakes

When we talk about evaluating ML models, we often jump straight to metrics. But that skips the real question: What kind of mistakes can your system afford to make?

The ML Process: From Raw Data to Deployed Model

Once you get past the hype and framework wars, most ML work reduces to a surprisingly small loop: problem definition, data acquisition, feature prep, training, evaluation, and deployment.

Supervised Learning

Nothing published under Supervised Learning yet.

Unsupervised Learning

Nothing published under Unsupervised Learning yet.

The notebooks and datasets that sit behind these entries live in my public repos ml-classics-notebooks (for supervised/unsupervised ML) and deep-learning-systems-notebooks (for neural networks and deep learning systems). I’ll link directly to specific notebooks as they stabilise.