The evolution of GitLab from Git repository to AI-native DevOps

Scaling through transparency and open core

GitLab started as a software project in 2011 and became a corporation in 2014 when Sytse Sijbrandij and Dmytro Zaporozhets launched it. Sijbrandij once cold-pitched a submarine inventor before working at GitLab. He and Zaporozhets decided to grant 10x voting shares to maintain control before going public. The company scales to 30 million users using a remote-first model supported by a 2,000-page public handbook. This handbook serves as the cultural backbone and a recruiting tool for the company. GitLab operates on an open core model, which includes a free Community Edition under the MIT license and a proprietary Enterprise Edition. This dual-licensing strategy targets organizations that require advanced features for large scale operations. While GitHub remains part of the Microsoft ecosystem, GitLab provides both cloud-hosted and self-managed deployment options. This flexibility allows teams to host the platform on their own infrastructure. I find the decision to decline a $10 million buyout early on telling, as it forced the company to pursue an IPO. The company prefers boring solutions by default to avoid unnecessary complexity, and decision-making remains decentralized. I would skip any platform that lacks such transparent documentation for its internal processes. Customer results sit at the top of the company values hierarchy.

Automating the software lifecycle

Integrated CI/CD workflows changed how teams manage software delivery. Continuous integration involves merging code changes into a shared repository and automatically testing each change. Continuous delivery prepares tested code so it is always ready for deployment, while continuous deployment automates the release to production. GitLab CI/CD uses .gitlab-ci.yml files to define these automated workflows through a declarative syntax. This integration provides a single platform for version control, testing, and deployment workflows. In 2018, GitLab 11.0 introduced Auto DevOps to automate the delivery process to Kubernetes clusters. The platform also includes support for Kaniko, which allows for secure image building in unprivileged containers. In 11.2, GitLab added JUnit test results in merge requests and interactive web terminals to help developers troubleshoot failed jobs. The parallel attribute was introduced in 11.5 to help teams speed up testing processes. Version 11.4 added the ability to use only/except rules in the .gitlab-ci.yml file based on modified paths, which helps with monorepo use cases. In 11.5, GitLab enabled deploying Knative to Kubernetes clusters with a single click. GitLab Runner provides a distributed architecture that supports virtual machines, containers, and various Kubernetes clusters. Security remains a central part of the DevSecOps ecosystem. Version 16.4 introduced browser-based Dynamic Application Security Testing to find vulnerabilities in modern web applications. The platform also includes dependency scanning and static analysis tools. I noticed one issue where the platform incorrectly duplicated SAST findings, which caused scan result policies to fail when different vulnerabilities appeared identical to the system.

The era of AI-assisted coding

The arrival of AI-driven development defines the recent history of GitLab Duo. This suite includes 15 AI-assisted features for all stages of the development lifecycle. Code Suggestions reached general availability in version 16.7 to help developers with code completion and generation. The tool supports 15 languages, including C++, C#, Go, Java, JavaScript, Python, PHP, Ruby, Rust, Scala, Kotlin, and TypeScript. I see no reason to worry about data privacy because GitLab does not use customer code to train its AI models. The 2023 State of AI in Software Development report noted that 83% of DevSecOps professionals believe AI implementation is necessary to avoid lagging behind in the industry. Version 18.4 introduced the GitLab Duo AI Catalog to let teams build and share custom agents. The Knowledge Graph provides a map of project files, routes, and references to help developers understand complex codebases. Version 19.2 expanded automation via GitLab Duo CLI and Custom Flows to replace manual multi-step workflows with agentic automations. Version 18.4 also included model selection, which gives customers control over which large language models power GitLab Duo. Features like Vulnerability Resolution empower teams to address vulnerabilities, while Discussion Summary helps condense lengthy conversations in Enterprise Agile Planning. By offering both cloud-hosted services and self-managed deployment options, GitLab provides a level of infrastructure control that Microsoft-owned GitHub does not match for teams requiring total sovereignty over their source code and deployment environments. Can organizations successfully manage the conflicts that arise between different automated AI agents?