Course with Bodil

Bodil — Reviewing AI Generated Code Without Missing Defects

Reviewing AI generated code, taught 1:1 by an AI tutor that plants the defects a model reliably produces. Four rounds of 200-line diffs for your tech leads, ending when they catch every planted high-severity defect twice running.

What you’ll master — Outcomes you can use

  1. A reviewer who catches the defects a model reliably produces
  2. They catch every planted high-severity defect twice consecutively

Your tutor

A staff engineer who has reviewed machine-authored diffs since the first Copilot rollout.

Terse · Specific · Never sarcastic

The plan — What’s inside

  1. Hallucinated APIs And Invented Config Keys
  2. The Edge Case The Model Skipped
  3. Silent Security Regressions
  4. Tests That Assert Nothing

4 parts

Questions about this course

How do I train my team to review AI generated pull requests properly?

Reviewing AI generated code, taught 1:1 by an AI tutor that plants the defects a model reliably produces. Four rounds of 200-line diffs for your tech leads, ending when they catch every planted high-severity defect twice running.

All Coding & Data templates · All templates