Generative AI
BasicAdvancedHow generative models produce new text — prediction and attention, prompting, retrieval, evaluation, adaptation, and what it costs to run them.
6
modules
16
lessons
5
runnable exercises
Sign in to begin →Sign in, then straight into the first lesson.
Course contents
01How generation worksBasic
How generation works
Generation as repeated next-piece prediction.
Tokens and sampling
The units models read, and how the next one is chosen.
Inside the model: attention
What the architecture actually does between input and next-token probabilities.
02Working with generative modelsBasic
Prompting and context
The context window, and why the whole input shapes the output.
Prompt patterns that work
Few-shot examples, step-by-step reasoning, and role framing — and when each fails.
Structured output
Getting machine-readable data out of a system that produces text.
03Meaning and retrievalBasic
Embeddings
Turning meaning into vectors so similarity can be measured.
Retrieval-augmented generation
Give the model the facts at question time instead of hoping it memorised them.
04Getting it rightAdvanced
Hallucination and grounding
Why fabrication is structural, and what actually reduces it.
Evaluating generated output
Scoring work that has no single correct answer.
05Adapting modelsAdvanced
Prompting, retrieval or fine-tuning
Three ways to make a model fit your problem, and how to choose.
Instruction tuning and alignment
How a next-token predictor became something that follows instructions.
Beyond text
Images, audio and mixed inputs — the same idea in a different space.
06Running it in productionAdvanced
Cost, latency and context economics
What you are actually paying for, and where the time goes.
Prompt injection and safety
What goes wrong when untrusted text reaches the context window.
Hosted or self-hosted
Where the model runs, and what that decides.
Ready when you are.
Opening a lesson takes an account. Nothing to install.