Introduction
What this course covers and how to read it
This is a from-scratch course on how AIAI (Artificial Intelligence)AI is the field of building systems that perform tasks normally requiring human intelligence — reasoning, perception, language, and decision-making. actually works, structured like a docs site rather than a textbook. Each lesson builds on the last:
- History & Landscape — how the field got here
- ML Fundamentals — learning as optimization
- Probability & Statistics Foundations — the math loss functionsLoss FunctionA loss function is a single number measuring how wrong a model's predictions are, which gradient descent minimizes during training. are built on
- Neural Networks & Backprop — the algorithm behind training
- Tooling & The Dev Stack — languages, frameworks, what you'd actually touch
- Computer Vision — convolutions and CNNsCNN (Convolutional Neural Network)A CNN is a neural network built around the convolution operation, which encodes locality and translation invariance for processing images efficiently.
- Attention & Transformers — the architecture behind modern AI
- Generative Models — GANsGAN (Generative Adversarial Network)A GAN trains a generator and a discriminator against each other, the generator learning to fool the discriminator into mistaking its output for real data., VAEsVAE (Variational Autoencoder)A VAE encodes inputs into a distribution over a compact latent space rather than a fixed point, so sampling from that space and decoding produces plausible new outputs., diffusion — generating new content
- LLMs — tokenizationTokenizationTokenization converts raw text into a sequence of integers a model can process, typically via subword schemes like byte-pair encoding (BPE)., pretrainingPretrainingPretraining is self-supervised training of a base LLM on massive amounts of text to predict the next token, the primary source of its knowledge and language ability., fine-tuningFine-tuningFine-tuning continues training a pretrained model on a smaller, curated dataset to teach it a specific behavior, such as following instructions., RLHFRLHF (Reinforcement Learning from Human Feedback)RLHF trains an LLM to match human preferences by learning a reward model from ranked response comparisons, then optimizing the LLM against that reward.
- Reinforcement Learning — learning from rewardRewardA reward is the scalar feedback signal a reinforcement-learning agent receives after taking an action — the only learning signal it gets, and often a delayed one., the mechanism behind RLHF
- Inference & Serving — what happens when you hit "send"
- Evaluation & Benchmarks — how models are actually scored
- Prompt Engineering — shaping behavior for free, no training involved
- RAG & Vector Databases — giving a model information it wasn't trained on
- Agents & Tool Use — giving a model the ability to act
- AI Security — prompt injectionPrompt InjectionPrompt injection is an attack where text an LLM processes — user input, a retrieved document, or a tool's output — contains instructions that override the application's intended behavior., jailbreaksJailbreakA jailbreak is a prompt crafted to bypass a model's own safety training and get it to produce output it was tuned to refuse., and defending a deployed system from misuse
- Applied & Agentic Systems — how the patterns above combine
Who this is for
You should be comfortable with basic programming and high-school math (functions, vectors, derivatives at an intuitive level). No prior MLMachine Learning (ML)Machine learning is the practice of writing programs that learn a function from data, via a loss function and an optimizer, instead of following hand-written rules. background is assumed — each lesson introduces its own vocabulary before using it.
How to read it
Lessons are meant to be read in order — later ones assume the vocabulary and intuition built in earlier ones.
Three things to know about the format:
- Technical terms link to the reference, a glossary of short standalone definitions. Follow them when a word is new; skip them when it isn't.
- Interactive widgets let you manipulate the thing being described — drag a learning rate around, step through backprop, watch attentionAttention (Self-Attention)Attention is a mechanism letting each position in a sequence weigh every other position via learned Query/Key/Value vectors, forming the core of the transformer. weights change.
- Deep dives are collapsed by default. Anything marked Under the hood contains the formal math and edge cases. The main text is complete without them, so ignore them on a first pass and open them on a second. The toggle at the top of each lesson opens them all at once.