Interviews Vector
Back to Roadmap
9
12 lessons

Reinforcement Learning

The foundation of RLHF and game-playing AI.

01

MDPs, States, Actions & Rewards

Learn
Python

A Markov Decision Process is five things: states, actions, transitions, rewards, a discount. Everything in RL — Q-learning, PPO, DPO, GRPO — optimizes over this shape. Learn it …

02

Dynamic Programming

Build
Python

Dynamic programming is RL with cheating. You already know the transition and reward functions; you just iterate the Bellman equation until `V` or `π` stops moving. It is the ben…

03

Monte Carlo Methods

Build
Python

Dynamic programming needs a model. Monte Carlo needs nothing but episodes. Run the policy, watch the returns, average them. The simplest idea in RL — and the one that unlocks ev…

04

Q-Learning, SARSA

Build
Python

Monte Carlo waits until the episode ends. TD updates after every step by bootstrapping the next value estimate. Q-learning is off-policy and optimistic; SARSA is on-policy and c…

05

Deep Q-Networks (DQN)

Build
Python

2013: Mnih trained one Q-learning network on raw pixels, beat every classical RL agent on seven Atari games. 2015: extended to 49 games, published in Nature, sparked the deep-RL…

06

Policy Gradients — REINFORCE

Build
Python

Stop estimating value. Parameterize the policy directly, compute the gradient of expected return, step uphill. Williams (1992) wrote it in one theorem. It is why PPO, GRPO, and …

07

Actor-Critic — A2C, A3C

Build
Python

REINFORCE is noisy. Add a critic that learns `V̂(s)`, subtract it from the return, and you get an advantage that has the same expectation but far lower variance. That is actor-c…

08

PPO

Build
Python

A2C throws away each rollout after one update. PPO wraps the policy gradient in a clipped importance ratio so you can do 10+ epochs on the same data without the policy exploding…

09

Reward Modeling & RLHF

Build
Python

Humans cannot write a reward function for "good assistant response," but they can compare two responses and pick the better one. Fit a reward model to those comparisons, then RL…

10

Multi-Agent RL

Build
Python

Single-agent RL assumes the environment is stationary. Put two learning agents in the same world and that assumption breaks: each agent is part of the other's environment, and b…

11

Sim-to-Real Transfer

Build
Python

A policy trained in a simulator that fails on hardware is a policy that memorized the simulator. Domain randomization, domain adaptation, and system identification are the three…

12

RL for Games

Build
Python

1992: TD-Gammon beat human champions at backgammon with pure TD. 2016: AlphaGo beat Lee Sedol. 2017: AlphaZero dominated chess, shogi, and Go from scratch. 2024: DeepSeek-R1 pro…