Deep Learning from Scratch

The deepest foundation for understanding every AI system. Start with the math, build a mini PyTorch, implement transformers from scratch, and understand exactly how modern AI works under the hood. For engineers who want to understand AI at the deepest level — not just use APIs.

The route

1. Math for AI & ML (full course) — Math for AI & ML. Linear algebra, calculus, probability, and optimization — all built in code, not on paper. The mathematical foundation for understanding every ML algorithm and neural network architecture.
2. Core ML: Algorithms from Scratch (full course) — Core ML: Algorithms from Scratch. Build every classic ML algorithm by hand — linear regression, logistic regression, decision trees, random forests, k-means, PCA — and verify against scikit-learn. Understand what the algorithms do, not just how to call them.
3. Deep Learning from Scratch (full course) — Deep Learning from Scratch. Build a mini PyTorch — autograd engine, layers, optimizers, training loops. Then implement CNNs, RNNs, and transformers from scratch. The course that makes you understand every AI system at the deepest level.
4. Transformers & LLMs from Scratch (full course) — Transformers & LLMs from Scratch. Implement attention, tokenization, positional encoding, and a full GPT-style transformer from scratch. Pre-train, fine-tune, and understand exactly how LLMs work — from the math to the production system.

What you build

A mini PyTorch with autograd, a from-scratch transformer that generates text, and the mathematical understanding to debug any AI system at the deepest level.

Every learning path

Create a free account — the opening phases of 24 of 30 courses are free, no credit card. Or see Pro pricing.

All courses · Pricing · About · FAQ · Glossary