Phase 09

Reinforcement Learning

Agents that learn by doing. The foundation of RLHF.

01MDPs, States, Actions & RewardsLearn~45 minutes 02Dynamic Programming — Policy Iteration & Value IterationBuild~75 minutes 03Monte Carlo Methods — Learning from Complete EpisodesBuild~75 minutes 04Temporal Difference — Q-Learning & SARSABuild~75 minutes 05Deep Q-Networks (DQN)Build~75 minutes 06Policy Gradient — REINFORCE from ScratchBuild~75 minutes 07Actor-Critic — A2C and A3CBuild~75 minutes 08Proximal Policy Optimization (PPO)Build~75 minutes 09Reward Modeling & RLHFBuild~45 minutes 10Multi-Agent RLBuild~45 minutes 11Sim-to-Real TransferLearn~45 minutes 12RL for Games — AlphaZero, MuZero, and the LLM-Reasoning EraBuild~120 minutes 13GRPO and Reinforcement Learning from Verifiable RewardsBuild~70 minutesQuiz

“Stay hungry, stay foolish.”

— Steve Jobs