Courses I recommend

The course that finally makes CSS feel deterministic instead of guesswork — layout, Flexbox, Grid and animation from first principles.

• CSS

The clearest path from 'I use TypeScript' to actually understanding the type system — the free tiers alone are worth it.

• TypeScript

Dense and demanding, but it rewires how you think about hooks, state, and React performance.

• React

The friendliest on-ramp to React — interactive, visual, and genuinely fun. Start here if it's your first framework.

• React

Closures, the call stack and the event loop, finally made concrete — the mental model I wish I'd started with.

• JavaScript

Free and official — the quickest way to get productive with the App Router by actually shipping something.

• Next.js

Free, focused CSS from someone who clearly loves it — perfect for building fundamentals that stick.

• CSS

The one course I point beginners to — zero to genuinely comfortable with the language, project-driven and kept current every year.

• JavaScript

An exercise-driven walk through routing, server actions and caching that covers both the App Router and the older pages directory — Kyle explains the why, not just the how.

• Next.js

A complete, free, project-based path into full-stack. Do the work and it genuinely takes you far.

• Node.js
• SQL

Scott Moss builds a real REST API in Node end to end — routing, auth and database, no hand-waving.

• Node.js

How to test with intent — what's worth testing, and how to write tests that survive a refactor.

• Testing

Learn SQL by writing real queries in the browser — the fastest way to stop fearing databases.

• SQL

The open-source reference for system design — diagrams, trade-offs and worked examples for interviews and real scaling.

• Architecture

A free, well-structured GraphQL intro from the fundamentals up, with tracks for the stack you actually use.

• GraphQL
• API

A free, one-hour intro to prompting from Andrew Ng and OpenAI — the fastest on-ramp to building with LLMs.

• LLM

MCP is quickly becoming the standard way to give models tools and data — this is the clearest explanation of how it actually works.

• MCP
• Claude

The famous code-first, top-down deep learning course — you build working models before the theory.

• Python
• ML