// Learn AI
An AI course from zero to the frontier
Pick a route for who you are, then follow 10 stages · 43 units. Every unit lists the algorithms, must-read papers, a project and a self-test.
43 unitsabout 385 hours in total223 algorithms with papers112 code & exercise links13 lessons written
01Choose a route
Each route picks the units you need, in order, with total hours and a weekly pace.
Everyday UsersNo code: in 2–3 weeks, learn to understand and use AI well and spot hallucinations and risks. 7 units~40h3–4 hours a weekApplication DevelopersPython is enough: in 3 months, ship a production-ready RAG / agent app you can evaluate and deploy. 16 units~126h8–10 hours a weekAlgorithm ResearchersFrom maths to frontier papers: in 6–12 months, read and reproduce papers on LLMs, alignment and reasoning models. 24 units~257h15+ hours a weekProduct Managers & LeadersUnderstand technical limits, cost structure and risk well enough to choose tools, start projects and set governance. 12 units~84h3–5 hours a weekDomain ScientistsWith a background in your own field, learn to do research with ML and LLMs — linked with the “AI × Disciplines” section. 15 units~143h6–8 hours a week
0210 stages
From AI literacy to the research frontier. Each step builds on the one before.
STAGE 0Prep: AI Literacy3 units · ~8hSTAGE 1Math & Programming Foundations5 units · ~61hSTAGE 2Classical Machine Learning4 units · ~32hSTAGE 3Deep Learning4 units · ~44hSTAGE 4Transformers & LLMs5 units · ~54hSTAGE 5Post-training, Alignment & Reasoning Models3 units · ~32hSTAGE 6Inference, Deployment & Efficiency5 units · ~36hSTAGE 7Multimodal & Generative AI4 units · ~38hSTAGE 8RAG, Agents & AI Engineering4 units · ~36hSTAGE 9Safety, Evaluation & Frontier6 units · ~44h
03Algorithms Atlas
Open the atlas →223 algorithms in 20 areas, each with its original paper, prerequisites and resources.
223 ALGORITHMSLook up an algorithmSearch by name, filter by area, difficulty and era.Open → MAPRelationship mapSee what to learn first, and how the field grew year by year.Open → PATH · 24 ALGORITHMSZero to Deep LearningFrom linear regression all the way to ResNet and the precursors of the Transformer — build intuition for how models are trained.Start →PATH · 19 ALGORITHMSUnderstanding an LLM End to EndFrom tokenisation to the Transformer, modern architectures, pre-training and inference — every part of an LLM, in nanoGPT order.Start →
04AI papers
Paper radar →New papers every day, a searchable archive and the essentials in reading order.
307 · 30 DAYSLatest papersDaily picks from arXiv and Hugging Face, each with a one-line takeaway.Open → 311 PAPERSArchive · searchNever removed; search by date, topic, discipline and keyword.Open → 150 ESSENTIALSEssential papersFrom backpropagation to reasoning models, in reading order.Open → 2026-W42Weekly top papersThe week’s most-upvoted papers, frozen every Monday. 5 weeks so far.Open →
05Hands-on: code, exercises & tutorials
Every unit links 2–3 free places to practise: browser demos, exercise sets, runnable notebooks and reference code.
STAGE 1Math & PyTorch drillsNumPy-100, Tensor Puzzles, the 60-minute PyTorch blitz.Start → STAGE 4Build a GPTTransformer Explainer, LLM Visualization, nanoGPT.Start → STAGE 5Fine-tune & RL a modelLoRA with Unsloth, DPO and GRPO with TRL, TinyZero.Start → 346 TUTORIALSCurated external tutorialsOfficial and university tutorials, filter by level and language.Open →
06Lessons you can read now
Full lessons already written. The course units link to them.
ⓘ As aulas completas estão disponíveis por enquanto em inglês e chinês simplificado.
- 01 What Is a Large Language Model? A plain-language explanation of LLMs: next-token prediction, parameters, training and limits. Iniciante 15 min ✓
- 02 How Machines Learn: Loss Functions and Gradient Descent The core loop of all modern AI training explained with one tiny example. Iniciante 18 min ✓
- 03 Neural Networks and Backpropagation Neurons, layers, activations and how errors flow backwards to update millions of weights. Intermediário 20 min ✓
- 04 Tokens and Embeddings: How Text Becomes Numbers Tokenisation (BPE), vocabularies and the vector space where meaning lives. Intermediário 15 min ✓
- 05 Transformers and Attention, Explained Self-attention, queries/keys/values, multi-head attention and the full decoder block. Intermediário 25 min ✓
- 06 How LLMs Learn: Pre-training, Fine-tuning and RLHF The three-stage recipe behind ChatGPT-style assistants, and newer methods like DPO. Intermediário 25 min ✓
- 07 What Do LLMs Learn Inside? Representations and Interpretability Features, superposition, circuits and world models — what we know and what we don't. Avançado 25 min ✓
- 08 How LLMs Generate Text: Decoding, KV Cache and Quantisation Sampling, temperature, the KV cache, batching and quantisation in one lesson. Intermediário 20 min ✓
- 09 RAG Explained: Retrieval-Augmented Generation Chunking, embeddings, vector search, re-ranking and grounded answers with citations. Intermediário 25 min ✓
- 10 How AI Agents Work: The Agent Loop Model + tools + loop: observe, think, act, and the patterns that make agents reliable. Intermediário 25 min ✓
- 11 Tool Calling and MCP (Model Context Protocol) How function calling works and how MCP standardises tools for every AI app. Intermediário 25 min ✓
- 12 How to Evaluate LLMs: Benchmarks, Arenas and Contamination What MMLU, SWE-bench and Chatbot Arena measure, and how to read leaderboards critically. Intermediário 20 min ✓
- 13 Prompt Injection and Tool Poisoning Why agents that read untrusted content can be hijacked, and the defences that work. Intermediário 20 min ✓
For AI and developers
- Todo o currículo está disponível em JSON com URL, status, público, software relacionado e fontes de cada aula.
- /api/learn.json · /api/learn/algorithms.json · /llms.txt