Intermediate8 weeks · Cohort-based
Machine Learning Foundations
How models actually learn — data, training, evaluation, and the maths that matters. The longest course we run, and the one that makes every other one click into place.
What you will be able to do
- Train, evaluate and debug a model end to end
- Read a results table and know what you are looking at
- Tell overfitting from a genuinely good result
- Explain how a transformer works to someone who asks
- Take a trained model to something other people can call
Course outline
- 01What learning from data actually means, and the maths you genuinely need.
- 02Data: collecting it, cleaning it, and the leaks that quietly ruin a model.
- 03Regression and classification, built by hand before reaching for a library.
- 04Evaluation — the metrics that matter and the ones that flatter.
- 05Overfitting and regularisation: why your model looked great until it did not.
- 06Neural networks: layers, gradients, and the training loop.
- 07Transformers, and what changed when attention arrived.
- 08Off your laptop: serving, monitoring, and knowing when to retrain.