Reinforcement Learning - A Learning Journey

Welcome! This site is a hands-on, code-first guide to Reinforcement Learning. Every concept comes with runnable Python, real-world context, and just enough math to know what's actually going on.

Whether you're here to understand Q-learning, train a PPO agent, or figure out why your reward curve looks like a heart-rate monitor during a horror movie - you're in the right place.

How to use this course

1. Start with Module 00: Setup & Tools to get your environment running.
2. Follow the modules in order - each one builds on the last.
3. Run the code. Break the code. Fix the code. That's where learning happens.
4. Check the boxes on exercises as you complete them - your progress is saved in your browser.

Course Modules

Module	Topic
00	Introduction
01	Math Foundations for RL
02	Multi-Armed Bandits & Exploration
03	MDPs, Bellman Equations & Dynamic Programming
04	Monte Carlo & Temporal Difference Methods
05	Function Approximation & DQN
06	Policy Gradients & Actor-Critic
07	Modern Deep RL (PPO, SAC, TD3)
08	Reward Design, Debugging & Diagnostics
09	Partial Observability & Generalization
10	Model-Based RL & Planning
11	RLHF & LLM Alignment
12	Multi-Agent RL
13	Sim-to-Real & Production RL
14	Paper Reproduction Project
15	Capstone

What you'll need

• Python 3.12+
• Curiosity and a willingness to watch training curves for longer than is healthy

Code examples live in the GitHub repo.