Agenda

After the Code: Reimagining DevSecOps for Agentic Teams

DevOps Stage — 45 minutes DevOps Agentic AI

Code generation worked. That's the problem. Every team that adopted an AI coding assistant in the last 18 months is now living with the consequences: more merge requests than reviewers can meaningfully read, dependency graphs shifting faster than security can triage, tests that nobody updated, and a recurring question:  what does this change actually break? that no single tool can answer, yet. The bottleneck has moved. It's no longer generating code; it's everything that happens to the code afterwards.This talk is a working tour of what comes next - agentic software delivery as a system of agents. We’ll follow code as it moves through review, security analysis, impact assessment across the codebase, and into the pipeline, with different tools doing the work: external coding agents such as Codex, Gemini, OpenCode, and Claude alongside GitLab’s Duo Agent Platform and Duo CLI, which helps connect those workflows inside a governed system.Along the way, we’ll examine questions I don’t think the industry has good answers to yet. What does review actually mean when most of the code wasn’t written by a human? What does governance look like when changes start proposing themselves? Where is an engineer’s attention best spent? You’ll leave with a picture of agentic DevSecOps as a working system and some questions worth asking your own team. Read More...

• Fatima Sarah Khalid [Senior Developer Advocate — GitLab]