Agentic AI | What if AI did your code review?
REXAI in production
2025-12-11 | 10:30 AM - 10:55 AM | Amphitheater Gaston Berger I Level -2
Information
What if an AI could do your code reviews for you?
In this session, we'll tell you how we created an intelligent agent integrated into GitLab that helps developers automatically review their code.
This agent uses open source artificial intelligence models like Mistral and LLaMA, which have been customized to understand good development practices (Clean Code, Craftsmanship, etc.).
With the n8n automation tool, we have set up workflows that allow the agent to:
• analyze Merge Requests,
• suggest improvements (naming, readability, complexity, etc.),
• propose or generate tests,
• and give a quality score to the code.
We will also talk about:
• the way in which the agent was designed (prompts, design, limits),
• what it has changed in our team (time saving, better code quality).
🎯 Why follow this talk?
Whether you're a developer, tech lead, DevOps, or simply curious about AI, this session will show you how AI can really help you in your daily work.
You will discover:
• how to automate smart code reviews without sacrificing quality,
• how to use open-source LLMs in a concrete setting, with simple tools like n8n,
• how to integrate AI into your GitLab workflows without changing your habits,
• and above all, how it can improve the quality of your code, save time, and boost your productivity.
A talk halfway between technique, practice and coding culture, well done – accessible, inspiring.
