Course 1 — AI & LLMs

Welcome to Course 1. This single course is split into three parts that build on each other. Read them in order, or jump straight where you need.

Level · Beginner → IntermediateDuration · ~7 hoursLessons · 18Language · English + Français

How the course is organised

Part 1 — History & concepts

A six-lesson tour that gives you the vocabulary and mental model to discuss modern AI without bluffing — from rule-based systems in the 1950s to today’s agentic AI.

1. History of AI — 70 years in 7 minutes.
2. From rules to data — the philosophical shift.
3. Types of machine learning — supervised, unsupervised, reinforcement.
4. The deep learning revolution — 2012, CNNs, the Transformer.
5. From GenAI to LLMs — what an LLM really is.
6. From LLMs to agents — prompting → RAG → agents → agentic AI.

Start Part 1 →

Part 2 — Supervised ML lifecycle

A six-lesson walkthrough of a real supervised ML project: from the business problem to production monitoring. Mostly what to do next and why, with minimal scikit-learn snippets at every step.

1. The big picture — the 17-step loop in one view.
2. Data: collect, explore, clean — the unsexy 80%.
3. Features: target, encoding, scaling — building X and y.
4. Train/test split & training — honest evaluation, one line of fit.
5. Evaluation & tuning — pick the right metric, search hyperparameters.
6. Deployment & monitoring — ship it and keep it alive.

Start Part 2 →

Part 3 — NLP basics

Six lessons climbing the full ladder of text representations — from raw strings to the embeddings that power modern LLMs. Each step is one tool you can drop into a real project.

1. What is NLP? — NLP, NLU, NLG and the classical pipeline.
2. Tokens & text normalization — raw string → countable units.
3. Stop words, stemming, lemmatization — three classical cleaning tools.
4. Bag-of-Words — the honest baseline that counts words.
5. TF-IDF — weighing words by how informative they are.
6. Embeddings & Transformers — from word2vec to the bridge to modern LLMs.

Start Part 3 →

---

How to follow this course

• Linear — Part 1, then 2, then 3. Recommended if you are new to AI.
• Bottom-up — Part 3 (NLP), then Part 1 (concepts), then Part 2 (lifecycle). For people with a programming background who learn better from concrete primitives.
• Just enough to talk — Part 1 alone is enough to hold a conversation about modern AI.

When you are done with this course, jump to Course 2 for the hands-on side: build a real Java code agent that runs entirely on your machine.