Senior AI Engineer

An overview of this role

As a Senior AI Engineer at Git Lab, you'll help build the foundation for Git Lab's transformation into an AI‑first company. Reporting to the Director, Enterprise AI, you'll be a hands‑on technical leader responsible for delivering internal AI‑powered solutions that drive measurable business outcomes.

Building fast matters, but it’s not enough on its own. This role starts with understanding the real problem: mapping how work moves across teams, tools, and handoffs, identifying the true constraint, and validating whether AI is the right solution before you begin development. From there, you’ll take ownership from discovery through deployment, combining strong engineering skills with systems thinking and business understanding.

Your initial focus will span Sales, Marketing, and Customer Support, where you will embed AI solutions into key systems and workflows. This role offers the opportunity to shape how Git Lab team members work, improve flow across the organization, and help advance our mission in a remote, asynchronous, and values‑driven environment.

• Diagnose business problems before building solutions. Map workflows, identify constraints, and confirm whether AI is the right intervention. Be prepared to say “this doesn’t need AI” when that’s the honest answer.
• Own AI initiatives end‑to‑end, from stakeholder discovery and technical design through implementation, deployment, and iteration.
• Design, develop, and ship AI‑powered solutions quickly, delivering working prototypes in days, not months, with a focus on practical outcomes and measurable business value.
• Improve organizational flow by building solutions that reduce bottlenecks, shorten lead times, and increase throughput. Measure success using flow metrics alongside adoption and ROI.
• Integrate AI capabilities into existing systems and workflows using APIs, orchestration tools, and modern AI platforms, including Git Lab Duo Agent Platform, where appropriate. The right tool wins, whether that’s custom code, a platform, or a well‑crafted prompt.
• Be Customer Zero: leverage and showcase Git Lab’s AI offerings wherever possible, feeding real‑world usage insights back to R&D.
• Partner closely with stakeholders across functions to understand the real constraints. Ask the right questions, bridge technical and non‑technical perspectives, and align on outcomes before jumping to solutions.
• Define and track success through business metrics, flow metrics, and feedback loops that make performance visible and actionable.
• Contribute to technical direction by evaluating tools, documenting patterns, and creating reusable foundations that help the team scale its impact.

• A Technologist at Heart – Genuinely invested in technology, the foundational and the cutting‑edge in equal measure. You’re as energized by a well‑designed API integration as you are by the latest foundation model release. You reach for the simplest solution that solves the problem well, rather than forcing new technology when proven approaches would do. AI is a powerful part of your toolkit, but it sits on top of solid engineering fundamentals, not in place of them.
• Competent, Confident Coding Skills – You can build working solutions end‑to‑end, write clean and maintainable code, and debug effectively. Whether your skills were honed in a traditional engineering role, through building automations, or shipping side projects, what matters is that you can deliver production‑quality work independently.
• AI & LLM Technical Depth – Strong proficiency in at least one modern scripting language (Python, JavaScript/Type Script, or similar) and a solid understanding of REST APIs, GraphQL, and integration patterns. Deep, practical experience with modern AI technologies, specifically: prompt engineering as a core discipline: designing effective system prompts, managing context windows, structuring multi‑turn interactions, evaluating output quality, and iterating systematically on prompt design.
• Model selection and cost‑performance trade‑offs – Understanding when a smaller fine‑tuned model outperforms a general‑purpose large one, when RAG is the right architecture…