Lessons Learned Building Developer Platforms
summary
A large legacy company builds an internal developer platform, spends about $10 million early on, and initially targets containerization across AWS, Azure and on‑prem, but only onboards 10–12 products per year against a 400 product portfolio. Strict security gating generates ~95% false positives and an estimated ~1,000,000 hours to remediate. Moving to a single cloud and embracing vanilla Kubernetes plus managed services and delamination (removing custom layers) plus a sourdough model for training finally accelerates migration and adoption.
AI-written summary · about
chapters (17)
- 0:00Developer platform lessons
- 1:25Naming pitfalls and origins
- 3:54Container platform journey
- 6:23Enablement team anti-pattern
- 8:53Security barriers and resistance
- 12:26Learning curves and capacity
- 15:32Adoption as work barrier
- 17:40Pivoting to legacy support
- 20:52Multi-cloud strategy
- 22:38Delaminating for simplicity
- 24:27Curse of dimensionality
- 27:15Needless optionality
- 30:04Rebuilding trust
- 32:54Year four transformation
- 34:40Key principles
- 36:27Q&A: Autonomy vs consistency
- 38:56Q&A: Delaminating and buy-in
related talks
26mLessons Learned from 7 years of Running Developer Platforms
59mLessons from Running Developer Platforms & Communities
32mFrom Cloud Naive to Cloud Native – Avoiding Mistakes Everyone Does
32mLessons Learned from 7 Years of Running Developer Platforms
33mWhy You Should Treat Your Platform as a Product
28mUnleashing Innovation by Unblocking your Developers
36mFrom Isolated Services to Integrated Ecosystems: Platform Evolution at Deutsche Bahn
11mBuilding Scalable Developer Platforms