Skip to main content
Skip to content
Menu

Codex Intermediate 3 parts Updated

Record a senior refactor walkthrough

You'll have a 5–10 minute screen recording showing how you use AI to tackle a real refactoring task in your codebase, narrating your prompts and how you validate the output.

What you'll make

Senior Refactor Walkthrough

  1. Setup: the refactoring target and why it's non-trivial
  2. Agent 1: primary refactor execution with live prompting
  3. Agent 2: adversarial review challenging the first agent's choices
  4. Conflict resolution: your judgment calls on competing recommendations
  5. Validation: testing the final result and what you'd do differently

Upload your 5-10 minute screen recording showing the complete multi-agent workflow on your actual codebase

The 3 parts

Each part ends with a check you can see.

  1. Map Your Refactor's Hidden Complexity

    Identify a real refactoring target and surface the non-obvious decisions that single-agent workflows typically miss.

    Done when

    You have a specific code target with at least 3 decision points where different approaches could be equally valid on the surface.

  2. Orchestrate Competing Agent Perspectives

    Set up two agents with different roles - one to execute the refactor, another to challenge it - and capture their conflict.

    Done when

    You have recordings of two distinct agent interactions: one focused on execution, one on adversarial review, with at least one clear disagreement to resolve.

  3. Resolve Conflicts and Validate Results

    Make judgment calls on agent disagreements, implement your resolution, and test whether the multi-agent process caught what single-agent would miss.

    Done when

    Your recording shows specific decisions on agent conflicts, implemented code that reflects those decisions, and narrated reflection on what multi-agent coordination revealed.

    Capstone: you put together your Senior Refactor Walkthrough.

Where this helps

An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.

About this lesson

What you'll be able to do

  • Record a 5-10 minute walkthrough showing how you coordinate multiple AI agents on a real refactoring task
  • Demonstrate at least two distinct agent roles (primary refactor + validation) with clear handoffs between them
  • Narrate your decision points for when to trust, challenge, or override multi-agent recommendations
  • Show your validation workflow for catching conflicts between different agents' suggestions

Who it's for

  • Senior developers who've used AI for coding but want to move beyond single-agent workflows
  • Technical leads responsible for code quality who need repeatable multi-agent review processes
  • Anyone comfortable with refactoring who wants to test whether orchestrated AI can catch what they'd miss

What it won't cover

  • Setting up specific agent tools or API configurations
  • Basic refactoring techniques - assumes you can already identify what needs refactoring
  • Production deployment considerations or team workflow integration

Skills you'll practise

  • Multi-agent task delegation for complex code changes
  • Cross-agent validation methodology
  • Conflict resolution between competing AI recommendations
  • Workflow narration for knowledge transfer