30 Days of RL

Your 30-day RL learning roadmap

How AI learns to reason, use tools, and complete real tasks. One practical lesson a day, in six stages.

Start with one idea

During reinforcement learning, a model makes attempts, receives reward-derived feedback, and is updated through a training algorithm. Simply retrying a task does not necessarily update the model.

The six stages

  1. 1Understand learningDays 1–6RL, pretraining, demonstrations, updates, agents and context
  2. 2Understand feedbackDays 7–12Rewards, preferences, checkable outcomes, PPO, GRPO and DPO
  3. 3Understand practiceDays 13–18Environments, rollouts, resets, verifiers, reward hacking and difficulty
  4. 4Understand capabilitiesDays 19–24Coding, tools, screens, conversations, long tasks and reasoning
  5. 5Judge the evidenceDays 25–28Generalization, setup failures, reliability and computational budgets
  6. 6Put it togetherDays 29–30An inspectable environment and the whole training pipeline

Your daily checklist

Each lesson is a short video. Work down the list as they come out.

Stage 1 · Days 1–6

Understand learning

  1. 01How AI learns without seeing the solution
  2. 02Three ways an AI model learns
  3. 03Show the answer or score the attempt?
  4. 04What actually changes when AI learns?
  5. 05How AI chooses its next action
  6. 06Did it learn, or read your correction?

Stage 2 · Days 7–12

Understand feedback

  1. 07What should we reward an AI for?
  2. 08Who decides which answer is better?
  3. 09When a test can grade the AI
  4. 10Why training updates need care
  5. 11Why AI tries the same question repeatedly
  6. 12Learning directly from preferences

Stage 3 · Days 13–18

Understand practice

  1. 13Where AI practises real tasks
  2. 14One task, different attempts
  3. 15Why every attempt needs a clean start
  4. 16The AI says done. How do we check?
  5. 17Full marks, wrong result
  6. 18Choosing tasks that teach something

Stage 4 · Days 19–24

Understand capabilities

  1. 19How coding agents practise fixing bugs
  2. 20Access to tools is only the beginning
  3. 21How AI finds the right button
  4. 22When the user changes their mind
  5. 23Which step caused the failure?
  6. 24Thinking longer versus training differently

Stage 5 · Days 25–28

Judge the evidence

  1. 25Skill or a leaked exam?
  2. 26Did the model fail, or did the setup break?
  3. 27One success versus dependable success
  4. 28What did that successful attempt cost?

Stage 6 · Days 29–30

Put it together

  1. 29Build a practice task you can inspect
  2. 30How an AI agent gets trained, end to end
After each episode, ask
  1. 1What is the task?
  2. 2What does the system observe?
  3. 3What can it do?
  4. 4How is success checked?
  5. 5Is this training or evaluation?

By the end, you should be able to explain the pipeline and inspect a small task setup. This introduction does not replace implementation practice or a full RL course.