Intermediate Backend Engineer (Ruby), AI Engineering, Custom Models

An Overview of This Role

The Custom Models team is responsible for Duo Self-Hosted a key component in GitLab AI that allows customers to run GitLab Duo features in completely private environments connecting GitLab to their own AI models.  The team works collaboratively with numerous teams to ensure a complete lifecycle of features that implement controls for customization and guarantee a smooth experience for our largest customers. In the near future the team will return to evaluating LLM models and advancing AI capabilities.

What You’ll Do

• Contribute to our features using JavaScript (30%) Ruby on Rails  (50%) and Python (20%) working with both frontend and backend components in a mature codebase.

• Collaborate closely with backend frontend product and UX teammates across time zones in an async-first environment.

• Write maintainable well-tested code and participate in code reviews that help you and your team grow.

• Maintain and improve the quality and stability of our applications by contributing to bug fixes and QA.

• Resolve production issues supporting customers and resolve production issues and learning along the way.

• Operate in an agile and fast-paced environment adapt to shifting priorities and project types and focus on outcomes.

• Take ownership of projects from proposal to production collaborating across teams to deliver end-to-end solutions.

What You’ll Bring

• Full-stack development experience: Proven ability to work across the stack with experience distributed approximately across JavaScript (30%) Ruby (50%) and Python (20%). While this is a backend role there will also be an opportunity to get involved with some frontend development.Comfortable working with both frontend and backend components in a mature codebase.

• JavaScript skills: Experience with modern JavaScript (and ideally a framework like Vue.js) particularly in the context of full-stack development.

• Strong Ruby and Python skills: Experience building and maintaining production applications using Ruby on Rails and  Python including work with background jobs APIs and data models.

• API development: Experience designing or extending REST and/or GraphQL APIs with attention to scalability maintainability and backward compatibility.

• LLM prompting and integrations: Hands-on experience with prompt engineering integrating LLMs into applications and building features powered by generative AI.

• SQL and data layer expertise : Strong command of writing efficient queries and working with relational databases such as PostgreSQL.

• Performance-minded: Skilled at identifying and resolving performance bottlenecks at both application and infrastructure levels.

Nice-to-Haves

• Familiarity with LLM evaluation RAG patterns or ML model deployment (especially using cloud platforms like AWS or GCP).

• Experience with Langchain LangSmith DSpy or similar LLM/AI frameworks.

• Contributions to open-source projects (especially GitLab or ML-related).

Why Join Us

This isn't just a job; it's your chance to shape the future of AI at GitLab.  You will be working on one of the most impactful products at GitLab. We operate asynchronously with a strong environment of support ownership and collective success. You'll gain deep expertise in Ruby Python and the GitLab platform while contributing to a core area of the product that directly supports millions of users.

As part of our team you’ll:

• Work on a widely-used open core product with passionate contributors and users

• Grow in a high-trust environment that values transparency and iteration

• Collaborate closely with Product and UX partners

• Have access to a defined career development framework and opportunities for mentorship and growth

About the Team

The Custom Models team consists of GitLab team members from around the globe. Engineers are primarily located across various European countries with some members in Australia and Canada.

The team works closely with these other teams within the organization:

• AI Framework

• MLOps

• Model Validation

• Duo Chat