
Record a 5-minute video walkthrough of your team's AI-assisted code review loop
You'll have a screen-recorded demo showing how Claude reviews your pull requests iteratively, catches errors in earlier feedback, and improves suggestions across multiple passes.
What you'll make
AI Code Review Loop Demo
- Open: The pull request and review setup (30 seconds)
- First pass: Claude's initial feedback (90 seconds)
- Second pass: Claude reviewing its own suggestions (90 seconds)
- Decision point: When to stop iterating (60 seconds)
- Close: What you learned about the loop (30 seconds)
Upload your screen recording with clear audio narration explaining each iteration decision
The 3 parts
Each part ends with a check you can see.
Benchmark Claude's Review Quality
Run Claude on 3 different PRs from your team, score its suggestions against your own review, and identify which feedback categories it gets consistently wrong
Done when
You have a scored comparison showing Claude's accuracy across 3 PRs with specific failure patterns documented
Build Your Iteration Rules
Test Claude's self-correction on your worst-performing cases from the baseline, then write specific rules for when to let it iterate vs when to override manually
Done when
You have documented stopping criteria that specify when to continue iteration vs when to intervene based on actual test results
Document the Process
Screen-record yourself running through the iterative review process on a fresh PR while narrating your intervention decisions using your tested rules
Done when
You have a 5-minute screen recording showing the iterative review process with clear narration of decision points and lessons learned
Capstone: you put together your AI Code Review Loop Demo.
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 working 5-minute demo of Claude iteratively improving your team's actual pull request review
- Document 3-5 specific cases where Claude's second-pass feedback caught errors in its first-pass suggestions
- Have a tested workflow for when to let Claude iterate vs when to intervene manually
Who it's for
- Development teams already using AI for code review who want to see if iterative loops improve quality
- Engineering managers evaluating whether to invest in AI-assisted review workflows
- Developers comfortable with pull requests and screen recording who want to optimize their AI review process
What it won't cover
- Setting up Claude API access or pull request tooling from scratch
- Comparing different AI models for code review
- Full CI/CD pipeline integration — this focuses on the human-AI review handoff
Skills you'll practise
- Iterative prompt refinement for code review
- Error pattern recognition in AI feedback loops
- Workflow documentation through screen recording
- Quality assessment of multi-pass AI suggestions