Home

Published

- 2 min read

Mojo, Modular & AI in CI/CD — Kicking Off a New Research Journey at FH Joanneum

Learning Programming
img of Mojo, Modular & AI in CI/CD — Kicking Off a New Research Journey at FH Joanneum

This summer, we’re kicking off a brand-new research project at FH Joanneum that blends three exciting areas:
🧠 AI-powered automation,
🛠 CI/CD pipelines, and
⚡️ cutting-edge tools like Mojo and Modular.com.

The goal? To explore how Large Language Models (LLMs) can be integrated into modern software development workflows to automatically generate tests, improve code quality, and accelerate DevOps.

🚀 Why This Research?

Writing tests is important—but let’s be real, it’s also repetitive and time-consuming. With the rise of LLMs like GPT-4 and open-source coding models, we’re now at a point where AI can assist in:

  • Writing unit and integration tests
  • Understanding code changes from commits or PRs
  • Increasing test coverage with minimal manual effort

Our research will focus on building a CI/CD pipeline that integrates this kind of automation—and seeing just how far we can push it.

💡 What’s Mojo and Why Does It Matter?

If you haven’t heard of Mojo, here’s the TL;DR:

It’s a new programming language from Modular.com that combines the simplicity of Python with the raw power of systems programming.

That means:

  • Native hardware acceleration
  • Compile-time performance
  • Python compatibility
  • Ideal for AI-heavy workloads

We’re not using Mojo yet—but we’re keeping a close eye on it as a potential tool for optimizing the more performance-critical parts of our future system.

🧪 What’s Planned for the Summer?

The project officially kicks off in summer 2025, and the first stage is all about:

  • Investigating how LLMs interpret and generate test cases
  • Designing a proof-of-concept CI/CD pipeline
  • Exploring how tools like Mojo or Modular’s inference engine could enhance speed or flexibility

We’re starting small, but the vision is big:
Imagine pushing code, and your pipeline instantly understands the changes, writes tests, and runs them—all powered by AI.

📬 Stay Tuned

This blog post is just the beginning. Over the coming months, we’ll share:

  • Technical deep dives
  • Experiments and findings
  • Lessons from integrating AI into testing workflows

Whether you’re a DevOps engineer, AI enthusiast, or just curious about the future of software development—follow along. This summer, we’re testing the limits of what AI can do in your toolchain.

Want updates? Let me know and I’ll help you turn this into a regular blog series or email newsletter.