Machine learning models · 60-lesson course

From model intuition to defensible ML systems

A practical adult-learning course covering the major machine-learning families, how they work, when to use them, how to evaluate them and how to operate them responsibly.

Course overview

This course moves from the core prediction loop through classical models, unsupervised learning, neural networks, foundation models, decision-making and real-world deployment.

How to use it: read one lesson, answer its retrieval check from memory, complete the practice task and explain the trade-off in your own words.

Start with Lesson 1

Foundations

Lesson	Focus
01	What a machine-learning model is
02	Learning settings and sources of supervision
03	Data, features, targets and loss
04	Train, validation and test sets
05	Bias, variance and generalisation
06	The maths needed for machine learning
07	Metrics, thresholds and calibration
08	Your first scikit-learn baseline

Classical supervised models

Lesson	Focus
09	Linear regression
10	Ridge, Lasso and Elastic Net
11	Logistic regression and generalised linear models
12	k-nearest neighbours
13	Naive Bayes
14	Decision trees
15	Random forests and extremely randomised trees
16	Gradient boosting
17	Support vector machines and kernels
18	Ensembles, calibration and interpretation

Unsupervised and probabilistic models

Lesson	Focus
19	k-means clustering
20	Gaussian mixture models
21	Hierarchical clustering
22	DBSCAN and density-based clustering
23	Principal component analysis
24	ICA and non-negative matrix factorisation
25	Anomaly and novelty detection
26	Bayesian modelling
27	Gaussian processes
28	Hidden Markov models and Kalman filters

Neural networks

Lesson	Focus
29	Perceptrons and multilayer perceptrons
30	Backpropagation and gradient descent
31	Activations, normalisation and regularisation
32	Convolutional neural networks
33	RNNs, LSTMs and GRUs
34	Embeddings and representation learning
35	Autoencoders and variational autoencoders
36	Graph neural networks

Generative and foundation models

Lesson	Focus
37	Attention and transformers
38	Large language models
39	Pretraining, prompting and in-context learning
40	Fine-tuning, LoRA, adapters and distillation
41	Retrieval-augmented generation versus training
42	Vision transformers and multimodal models
43	Generative adversarial networks
44	Diffusion models
45	Foundation-model evaluation and safety

Decision, time and structure

Lesson	Focus
46	Time-series forecasting
47	Recommender systems and ranking
48	Contextual bandits
49	Reinforcement learning and Markov decision processes
50	Q-learning, policy gradients and actor-critic methods
51	Causal inference and structural causal models
52	Survival analysis
53	Structured prediction and conditional random fields

Real-world machine learning

Lesson	Focus
54	Feature engineering and data leakage
55	Imbalanced data and rare events
56	Interpretability, fairness and calibration
57	Deployment, latency and cost
58	Monitoring, drift and retraining
59	Choosing the simplest model that works
60	Capstone: design, train and defend an ML system