AI & ML Foundations
Understand what is running underneath

The business arithmetic that decides pricing, discounting and capacity — taught in five short lessons and one worked case. You will learn why average cost is a report and marginal cost is a decision, what elasticity actually measures and why it changes at every price, where revenue peaks and why profit peaks somewhere else, and how to turn any proposed discount into the volume it silently assumes. Ends on a company with two customer segments, a nearly full factory and five decisions to make — where the winning plan sells fewer units than the losing one.

The front door of AI & ML Foundations, and the only path in it that assumes nothing at all. Start with the definition — why the word keeps moving, and the one meaning of it that can actually be scored. Then the map: every approach the field has tried, organised by a single question. Then the two that came first, rules and search, taken seriously and pushed until they break — which is the reason machine learning exists.

The second path of AI & ML Foundations, and the only one with no dataset and no maths in it — because everything here is a distinction, and a distinction is learned by watching one thing behave three ways. A robot with a bow, a target it cannot see, and three different things it can be told after each arrow. By the end you can look at a business question and say which kind of learning it needs, or that it needs none.

The first path in AI & ML Foundations where you build a model rather than sort problems, and it is a tree first because a tree's reasoning is a conversation you can have out loud. One question runs the whole path — which question do you ask first, and how do you know it was right — asked by feel on sixteen animals, by arithmetic on twenty-four films, and then on 1,500 real employees where feel fails. By the end you can build a tree by hand, read every decision it makes, and say when it has stopped learning and started memorising.

The first path in AI & ML Foundations. One dataset — a media budget split across TV, radio and newspaper in 200 markets — carried from the question through the model to the decision it feeds. By the end you can fit a regression, read what its coefficients claim in base and incremental sales, and tell the difference between a model that scores well and a model that is right.

The shortest path in AI & ML Foundations, and a bridge rather than an introduction. You have already built two models: a line fitted to 200 markets, and a tree grown on 24 films. This path asks the same four questions of both — what shape may the answer be, what is free, what is it scoring, how did it search — and finds that for the tree, two of them come back empty. Then it spends a lesson on each of those two: the cost somebody chose for you, and the search that stopped where it stopped.