• Validation that confirms agent-generated code does what was intended, so a passing suite means a working feature and not just a satisfied proxy.
• Mid-loop feedback that helps an agent self-correct toward working software instead of green assertions.
• Language-aware approaches under one shared model of what "trusted" means, starting with Rust, Ruby, and Go, while allowing the right technique to differ by language.
• Real measurement of whether engineers trust and adopt the path, fed straight back into the platform.

What you’ll do

• Own the technical thesis for how GitLab validates software when an agent is the author, and turn it into an architecture the team can build against.
• Prototype the hard parts by building early versions of feedback mechanisms that catch the gap between passing tests and working software, proving what works before the team scales it.
• Develop approaches for GitLab's primary stacks, starting with Rust, Ruby, and Go, under one shared definition of "trusted," while allowing the right technique to differ by language.
• Design the mid-loop feedback an agent needs to self-correct toward working software rather than green assertions, and build the systems that deliver it.
• Collaborate with engineering teams to adopt these approaches, measure adoption, engineer trust, and validation quality, and use those findings to improve the platform.
• Set technical direction the team can build on, review designs, and help teammates grow their technical judgment.

What you’ll bring

• Deep, current, hands-on experience building real software with artificial intelligence agents, ideally across more than one language, and a clear understanding of why agent output can pass tests yet fail in production.
• A strong point of view on the validation gap and concrete ideas for closing it, backed by approaches you've built or tried.
• The technical range to design systems rather than tune a single framework, and the judgment to know when a solution should be general or language-specific.
• You prototype to learn, and you ship.
• The ability to influence engineers across an organization without formal authority through working code and sound reasoning.
• Strong written communication in an all-remote, asynchronous environment.
• Alignment with GitLab's values, especially collaboration, iteration, and transparency.

This isn't a quality assurance or test automation role. We aren't looking for depth in a particular testing framework. We're looking for someone who has built with agents, experienced the validation gap firsthand, and wants to solve it.

About the team

The Trusted Agentic Development team is part of GitLab's Developer Experience group. We build platforms and practices that help GitLab engineers validate agent-generated software with confidence. We work across engineering teams and programming languages to understand how people use agents, test new approaches in real workflows, and create a shared model of trusted agentic development that teams choose to adopt.

Poland base salary range:272,000 PLN – 408,000 PLNThe base salary range for this role’s listed level is currently for residents of the EU, applicable in the specified country or countries only. Sales roles are also eligible for incentive pay targeted at up to 100% of the offered base salary. Grade level and salary ranges are determined through interviews and a review of education, experience, knowledge, skills, abilities of the applicant, equity with other team members, alignment with market data, and geographic location and having regard to any applicable collective agreements. The base salary range does not include bonuses, equity, or benefits.

Originally posted on Himalayas