Architecting AI-DLC for the organization
summary
Organization builds two agents (discovery/planning and engineering) plus a workspace harness that commits across multi-repos, creates PRs, and distributes skills via a Cursor plugin for 200 engineers; specs-first policy, inner source by default, and written context live in Confluence, Google Drive and Jira. Automated fitness functions score outputs with RAG results, auto-create Jira tickets and hand fixes to agents for PRs. The system already raises throughput and reuse but governance, context-rot prevention and cost controls remain unfinished.
AI-written summary · about
chapters (9)
- 0:00Introduction and context
- 1:34Defining the AI-DLC paradigm
- 5:57Organizational challenges
- 10:20Laying the foundation
- 13:51Development principles
- 15:22Phases of the lifecycle
- 17:14Mechanics and harnesses
- 22:24Fitness functions
- 25:43Observations and conclusion
related talks
51mBeyond Code Generation: Governing AI Agents Across the DevSecOps Lifecycle
31mModern Architecture for AI-Augmented Development Teams
31mEngineering the harness
35mThe Agentic AI Engineer
08mBuilding an AI Platform Your Engineers Actually Trust
30mArchitecting Integrity: Responsible Software Evolution in the AI Era
14mAgent-Driven Software Development
30mBuilding Tools with Tools: Modern AI-Driven Development