Refactoring Podcast AI Coding meets Code Health 🪄 — with Stuart Caborn
23 snips
Apr 17, 2026 
Stuart Caborn, distinguished engineer at loveholidays who builds platform engineering and AI-first developer workflows. He discusses applying AI to real-world software delivery, keeping elite code health while 60%+ of production code is AI-assisted. Short, concrete stories cover data products, guardrails, skills-as-executable-docs, and how they measure and scale safe AI-driven change.
AI Snips
Chapters
Transcript
Episode notes
Prioritize Hot Code For Faster Improvement
- Love Holidays improved code health before AI and focused on high-change "hot" areas rather than trying to fix the whole codebase at once.
- They asked engineers to keep frequently touched code high quality, treating new layers like lava that should solidify well for future changes.
Make Code Health A Visible Data Product
- Make code health and quality a visible data product with dashboards so teams can see and act on it.
- Love Holidays published code health as a data product and exposed it to everyone, then celebrated top teams to create friendly competition.
Incident Analysis Turned Into Reusable Prompt
- An engineer used the MCP in an incident to estimate booking loss and a trader corrected the model by replaying the question to the AI.
- They distilled the interaction into a reusable prompt so anyone could run instant incident impact analysis.