Backend Engineer, AI Engineering: Duo Chat

Job description

About this role

An overview of this role As a Backend Engineer on the Chat Engine team, you'll build the engine behind GitLab Duo Chat, the conversational AI experience for GitLab. You'll work primarily in Python to build and maintain our agentic runtime: the Flow Registry, LangGraph flows, and the Duo Workflow Service. You'll also work in the GitLab Rails monolith, where Chat connects with the product. You'll own scoped parts of the system, ship small features and improvements with minimal guidance, and collaborate with the team on larger projects. You'll work alongside senior and staff engineers who will partner with you on design and support your growth.

Responsibilities

• Develop secure, well-tested, and performant features and improvements for GitLab Duo Chat.
• Implement and extend agentic flows in Python using LangGraph, registered through the Flow Registry and served by the Duo Workflow Service.
• Integrate Chat capabilities with the GitLab Rails monolith, including its GraphQL application programming interface (API).
• Write maintainable code that meets our internal standards for a high-scale environment, including tests with every merge request.
• Review code within our Code Review Guidelines and respond promptly to contributions from people who don't work for GitLab.
• Identify technical debt and propose and implement solutions that improve the team's efficiency.
• Collaborate with Product Management, User Experience, and Frontend team members to improve product quality, security, and performance.
• Participate in Tier 2 or Tier 3 weekday, weekend, and occasional night on-call rotations to help troubleshoot product operations, security operations, and urgent engineering issues.

Qualifications and requirements

• Professional experience with Python, including building and testing production services.
• Working proficiency with Ruby on Rails, or a willingness to learn it, since a meaningful part of Chat lives in the GitLab monolith.
• Exposure to, or strong interest in, building on top of large language models, including prompting, tool or function calling, evaluation, and the failure modes of non-deterministic systems.
• Experience designing and building APIs, including diagnosing and optimizing performance issues, and comfort working with well-documented third-party services.
• A security-minded development approach that applies secure coding guidelines early in the development process.
• Clear and concise written communication about technical problems and iterative solutions, which is essential in a remote, largely asynchronous environment.
• Manage work independently, organize priorities, and ask for help when an issue is taking longer than expected.
• Demonstrate GitLab's values by collaborating effectively in a fully remote, intensely iterative environment.
• Experience with LangGraph, LangChain, or a comparable agent orchestration framework.
• Experience with GraphQL and RSpec.