Closing the Agentic DevOps Loop with Microsoft and Moderne
summary
GitHub Copilot runs as a cloud agent calling Moderne’s MCP and Azure MCP to make deterministic code changes across repos; a live Java 11 to Java 25 migration uses Moderne’s lossless semantic tree and deterministic recipes to modify 21 files, run build/tests and produce a PR, cutting token use by about 30,000 compared with an unguided run. A CVE sweep finds 17 issues and auto-remediates 11, stopping and surfacing the rest for human review.
AI-written summary · about
chapters (10)
- 0:00Introduction and speakers
- 1:10AI in development workflows
- 3:56Evolution of DevOps with AI
- 6:25Agentic DevOps lifecycle
- 9:17Modern's role in agentic DevOps
- 13:54Code as data for agents
- 20:55Live demo of agentic modernization
- 37:13Microsoft's AI platform integration
- 39:18Commercial models and billing
- 41:47Q&A and closing
related talks
27mThe State of Large-Scale Code Modernization
22mUnlocking AI Powered DevOps Within Your Organization
27mNavigating Application Modernisation - Leveraging Gen-AI
14mModernize .NET Apps
16mDriving agentic innovation w/ MCP as the backbone of tool-aware AI
31mMonorepos, Mainframes, and Modus Operandi Accepted
06h 22mUsing AI to Enhance Developer Productivity
03h 54m.NET Day of Agentic Modernization 2026
