GitLab Introduces Agent-Based DevSecOps with Vertex AI

GitLab, a provider of intelligent orchestration platforms for DevSecOps, has announced an expanded partnership with Google Cloud. This partnership allows Google Cloud users to leverage the GitLab Duo Agent Platform with their existing Vertex AI models. Furthermore, by centralizing agent inference from within GitLab to Vertex AI, Google Cloud users can now manage their GitLab Duo Agent Platform usage in line with their existing contracts and control frameworks on Google Cloud. This helps avoid duplicate spending and shadow IT that bypass procurement processes.

As AI agents become central to software development, a robust environment with both excellent models and strong governance is essential for them to be effective in enterprise settings. This partnership brings these two elements together. AI agents on the GitLab Duo Agent Platform will now have direct access to Vertex AI models on Google Cloud, allowing users from both companies to leverage both the expected model performance and GitLab’s governance.

GitLab functions as a centralized record system for managing issues, code, merge requests, pipelines, and security findings. When AI agents generate code suggestions, they draw context from the starting issue, existing patterns in the repository, and the CI pipeline used for verification; all of this processing is completed within GitLab. AI agent operations are performed within the same mechanisms as the access controls, approval rules, and audit logs that developers normally use, so governance is not compromised when the agent takes over tasks.

- Advertisement -

Also Read: AIR Is Not Selling AI Tools It Is Selling Execution

Vertex AI Model Garden offers Google Cloud and GitLab users a choice of foundational models, such as Gemini, allowing development teams to select a model based on performance, cost, and regulatory requirements. Furthermore, GitLab Self-Managed users can leverage GitLab’s Bring Your Own Model (BYOM) to position approved providers and gateways in line with their security model design. GitLab AI gateways, whether GitLab-managed or self-hosted, can run on Google Cloud runtime environments such as GKE and Cloud Run, enabling teams to run AI workloads on their existing infrastructure.

SOURCE: PRTimes