🧠 Learn AI the school way
How models actually work (tokens → embeddings → attention → RLHF & LoRA), then how to use them well (prompting, RAG, agents, diffusion) — every concept as a school story, every lesson with a numbered diagram and a hands-on lab. The labs are zero-dependency Python: no GPU, no API key, no pip install.
🎬 Companion video: 9 AI Concepts Explained in 7 minutes (ByteByteAI) — the aerial view; this course is the slow ground tour of every concept it names.
🔬 Part 1 — INSIDE the model
- rules are grown from examples, not written 📖
- text becomes puzzle pieces 🧩 seated by meaning 🗺️
- one game: guess the next piece 🎲 — with glancing 👀
- finishing school: etiquette, gold stars, sticky notes ⭐
🧰 Part 2 — USING it for real
- ask the librarian well; mind the small desk 🗣️🪑
- the confident kid makes things up — defenses 🙋
- open-book exams (RAG) 📖 and hall passes (agents) 📋
- un-blur the static: diffusion & multimodal 📺
🗺️ The big picture — one diagram, both worlds
The whole course on one canvas. Click for the 4K version.
🔬 Part 1 — inside the model (lessons 1–7)
One git branch = one idea; branch 05 contains lessons 01–05. Labs run with plain Python 3.
📖 What is AI
lesson-01-what-is-aiRead lesson →See the diagram ↗✏️ Training
lesson-02-trainingRead lesson →See the diagram ↗🧩 Tokens
lesson-03-tokensRead lesson →See the diagram ↗🗺️ Embeddings
lesson-04-embeddingsRead lesson →See the diagram ↗🎲 Next-token prediction
lesson-05-next-tokenRead lesson →See the diagram ↗👀 Attention
lesson-06-attentionRead lesson →See the diagram ↗⭐ RLHF & LoRA
lesson-07-rlhf-loraRead lesson →See the diagram ↗🧰 Part 2 — using it for real (lessons 8–12 + bonus 13)
No installs here either — the labs use any chatbot you already have, plus paper and honesty.
🗣️ Prompting & context
lesson-08-prompting-contextRead lesson →See the diagram ↗🙋 Hallucinations
lesson-09-hallucinationsRead lesson →See the diagram ↗📖 RAG
lesson-10-ragRead lesson →See the diagram ↗📋 Agents & tools
lesson-11-agentsRead lesson →See the diagram ↗📺 Diffusion & multimodal
lesson-12-diffusionRead lesson →See the diagram ↗🔌 Bonus: MCP
lesson-13-mcpRead lesson →See the diagram ↗# take the course locally:
git clone https://github.com/BaluRaut/learn-ai-school.git
cd learn-ai-school
python3 demo/bigram_model.py # a language model in 60 seconds
git checkout lesson-01-what-is-ai # then lesson by lesson📐 The lesson diagrams — follow the numbers
Every lesson as one numbered box-and-arrow diagram (purple = inside the model, teal = using it) — also on a standalone page.
1 📖 What is AI
Rulebooks vs learning from examples — the rules are grown, not written.
2 ✏️ Training
Practice tests + the red pen — loss and tiny downhill nudges.
3 🧩 Tokens
Cutting text into puzzle pieces — why 'strawberry' is hard.
4 🗺️ Embeddings
The seating chart of meanings — similar words sit together.
5 🎲 Next-token prediction
The sentence-completion game, with a temperature dial.
6 👀 Attention
Kids glancing around the class — who does 'it' refer to?
7 ⭐ RLHF & LoRA
Library, etiquette class, gold stars — and sticky notes.
8 🗣️ Prompting & context
How you ask the librarian — and the small desk it fits on.
9 🙋 Hallucinations
The confident kid who never says 'I don't know'.
10 📖 RAG
The open-book exam, automated with the seating chart.
11 📋 Agents & tools
A student with a to-do list and a hall pass.
12 📺 Diffusion & multimodal
Un-blurring TV static, step by step, toward your words.
13 🔌 Bonus: MCP
The universal plug — deep-dive course: learn-mcp-school 🔌