Optimizing the engineering value stream
summary
Shows how a focus on the engineering value stream (plan → code → validate → deploy → monitor) yields repeatable, reliable delivery, with concrete examples: Night Capital loses $440M from a bad deploy; use a platform mindset and local generative AI stacks like AMA and Fabric plus Open Web UI and Continue for private LLM coding and testing; prioritize work by value × pain × feasibility; measure with a small balanced set of metrics and apply ratchet metrics (e.g., raise code coverage from 30% to 35% as a new baseline) to make gains sustainable.
AI-written summary · about
chapters (10)
- 0:00Introduction and background
- 1:48Night Capital deployment disaster
- 3:14Engineering value stream explained
- 5:42Discipline over shortcuts in engineering
- 8:53Adopting generative AI safely
- 13:07Prioritizing engineering improvements
- 15:18Measuring engineering effectiveness
- 20:13Metrics and culture pitfalls
- 23:28Balanced approach to metrics
- 25:54Summary and key takeaways
related talks
15mOptimizing the engineering value stream
01h 00mDevOps – Value at Speed
31mEngineering Effectiveness: Achieving your goals with less
44mEnhancing the Performance of Engineering Teams
43mEngineering Platforms Evolved
17mSociotechnical Platform Engineering
51mImproving Software Flow
27mTurning Platform Engineering Work Into Business Value Leadership Understands