AI Foundations
BasicAdvancedHow machines learn from data — the paradigms, the training mechanism, the classic algorithms, neural networks, and what it takes to judge and run a model responsibly.
7
modules
20
lessons
15
runnable exercises
Sign in to begin →Sign in, then straight into the first lesson.
Course contents
01What AI isBasic
What is AI?
A working definition, and the shift from rules to learning.
Why AI works now
The idea is old; three things changed to make it practical.
Learning from data
How a system improves at a task by learning from examples.
02How machines learnBasic
Supervised vs unsupervised
Two styles of learning, depending on whether data is labelled.
Regression and classification
The two supervised tasks, and why the difference changes everything downstream.
Features and data quality
What you feed a model decides more than which model you pick.
Learning from consequences
The third paradigm: no labels, just rewards for what worked.
03Training a modelBasic
Loss and gradient descent
The mechanism by which a model actually improves.
Overfitting and generalisation
Memorising the training data is easy; the goal is something else.
Train, validation and test
Three sets, three jobs, and the one you must not touch.
04Judging a modelBasic
Evaluating a model
Measure how often a model is right — and why accuracy can mislead.
Precision, recall and the confusion matrix
Four numbers that say what kind of wrong a model is.
05Classic algorithmsAdvanced
Linear and logistic regression
The two workhorses, and why simple models are still the default.
Decision trees and ensembles
Models that ask questions — and why a forest beats one tree.
Clustering and nearest neighbours
Two ideas built on distance, and the scaling trap they share.
06Neural networksAdvanced
How a neural network works
Layers of weighted sums, and the one ingredient that makes depth matter.
Training deep models in practice
Batches, epochs, and the settings that decide whether training works at all.
07Models in the worldAdvanced
Bias and fairness
Where unfairness enters, and why deleting a column does not remove it.
Interpretability
Knowing why a model decided, and what an explanation does not tell you.
Deployment and drift
A model is not finished when it is accurate; that is when the work starts.
Ready when you are.
Opening a lesson takes an account. Nothing to install.