Scaling the Mountain: A Framework for Tackling Large-Scale Tech Debt
summary
Python-powered framework surfaces and tracks actual tech-debt items across a 3M-line monolith, runs incremental analyzers in CI, stores snapshots in S3, and generates leaderboards and Slack shoutouts; it runs codemods (built with lipcst and DeepCST), fixes repetitive issues automatically, and reports concrete wins — reduced 70,000+ pyright-ignore annotations and raised type coverage from ~60% to ~99.6%, plus a monolith-to-async conversion — with measurable counts, weights, and task links.
AI-written summary · about
chapters (16)
- 0:00Scaling large tech debt
- 1:28Outline and challenges
- 4:43Why large codebases struggle
- 5:27Measure tech debt systematically
- 8:20Tracking with analyzers and goals
- 10:34Recognize contributions and accountability
- 13:47Leaderboards for tech debt work
- 17:50Culture of accountability
- 18:32Team reports and planning
- 19:40Automate fixes with tools and AI
- 24:21Case study: ZIP tech debt reduction
- 26:51Summary of three strategies
- 27:14Q&A: Setup time and business impact
- 28:18Q&A: Financial benefits tracking
- 30:06Q&A: App logic tech debt
- 32:12Closing and discussions
related talks
36mNo More ‘Someone Else’s Problem’: Tackling Tech Debt at Scale
21mTeach Your Old Code New Tricks: Automating Code Quality in Large Projects
36mNo More "Someone Else’s Problem": Tackling Tech Debt at Scale
10mPaying down management debt without burning out
20mPrioritizing Technical Debt as if Time and Money Matters
48mA Crystal Ball to Prioritize Technical Debt
45mTechnical Debt Prioritisation - Identifying & Fixing Highest ROI Issues
30mPrioritizing Technical Debt as if Time and Money Matters