🦾 Robot Learning — The Self-Improving Lab

A track about systems that keep getting better by themselves: simulated robots practice, get scored, and the best version is promoted — day and night. Lesson by lesson we build one piece of the loop, from reward basics to a walking humanoid and a 24/7 dashboard.

📖 Sources: Sutton & Barto, Reinforcement Learning: An Introduction (free online) for the theory; AlphaGo Zero and Chip Huyen's Designing Machine Learning Systems for the industry loop.

♻️ The loop: collect experience → train → test vs the champion → deploy → monitor → collect again.

🖤 Each lesson opens with a blackboard slide deck before the quiz. See rl-01/course-goals.txt for the lesson plan, the full book roadmap, and the AI image prompts.

Lessons 📖

🎯 rl-01 Learning by Reward — theory & the loop

🎢 rl-02 NEW CartPole — your first real environment

🚶 rl-03 2D walker & evolution strategies

🦾 rl-04 Robot arm & PPO

🤖 rl-05 3D humanoid on a cloud GPU

🥋 rl-06 Motion imitation

🌪️ rl-07 Randomization & shoves

👁️ rl-08 Computer vision for control

♻️ rl-09 The 24/7 self-improving lab