Lessons Learned: The Why, What and How of Software Engineering Intelligence
summary
Claims a per-customer ML engine trained on historical SDLC data (five years cited) that correlates data from 35 tools to produce risk scores, sometimes as specific as change failure probability 5.7% or warnings approaching 80% risk. It demos a bubble “change weather” dashboard, exposes results as an API consumed by CI/CD to automate a traffic light system, and evaluates changes across 60–80 dimensions to cut CAB workload and gate deployments.
AI-written summary · about
chapters (11)
- 0:00Software engineering intelligence intro
- 3:14Why it matters
- 6:03DevOps and platform evolution
- 8:52Data silos problem
- 10:40Combining data sources
- 12:25Two abstraction layers
- 15:18Predictive AI training
- 18:09Automating change boards
- 24:10Live demo: change weather
- 29:49Risk prediction in CI/CD
- 32:21Q&A
related talks
16mEngineering Intelligence & Developer Productivity in the AI Era
17mLeadership in AI-assisted engineering
39mThe evolution of Devops in the age of AI
41mFrom Lab To Life: Practical AI System Evaluation
41mWhen Software Starts Thinking: The Now of Quality Engineering
22mLeveraging Quality Intelligence and AI to Improve the Quality of Java Applications
31mElevating DevOps with Integrating Analytics into CI/CD Pipelines
34mYou say you've got an "AI SDLC"