Code Forward: Aalto students join Codemate to build Generative UI

Every autumn, new software projects start quietly in classrooms and meeting rooms. This year, one of them looks a little different.

At Aalto University, a group of computer science students have teamed up with Codemate to work on a project centred on Generative User Interfaces (GenUI) and AI-assisted development. Over the next six months, they will design and build an application for real student needs, using Codemate’s own GenUI platform, Rebel AI Studio, as their foundation.

This blog series, Code Forward, follows that journey — the ideas, the detours, the surprises, and the lessons learned on both sides.

Aalto’s software project, with a twist

The collaboration takes place within Aalto University’s Software Project course, where teams develop a software system for a real client using the Scrum framework. Each team includes a master’s level student as Scrum Master, and the company acts as Product Owner.

In this project, Codemate fills the role of PO. The student team has eight third-year students and a more senior Scrum Master. From Codemate, mentors Ville Lindfors, Sole Piirainen and Iiro Pelttari bring experience from client work and from developing Rebel AI Studio itself.

The brief is simple on paper: build something that genuinely helps fellow Aalto University students by using Rebel AI Studio as the base.

*Ville Lindfors (far left) and the Aalto University students working at the Codemate Helsinki office. Picture: Sole Piirainen*

Why this collaboration matters

For Codemate, the project is both a partnership and a practical stress test of a platform already used in production. Students approach Rebel AI Studio with no legacy assumptions and no hesitation about pointing out unclear or clumsy workflows. That perspective can be difficult to access internally.

“Codemate uses Rebel AI Studio as part of our solution stack when it’s the right fit. Involving a student team gives us a fresh-eyed perspective to strengthen both the product and the developer experience,” says Ville Lindfors, Director of AI, Data & Cloud at Codemate. “It helps us validate what’s working well and identify where the user experience can be further refined.”

For the students, the value lies in contributing to a real product rather than a hypothetical assignment. They see how a professional team structures work, makes decisions under constraints, and uses AI in day-to-day development rather than in isolated exercises.

“I expected a top-down process where we’re told exactly what to do,” one student says. “Instead, we’re expected to come up with ideas ourselves, and that’s been a very pleasant surprise.”

Two-way learning

“We are very open about the fact that this is also a learning project for us,” says Ville Lindfors. “We want to see where Rebel AI Studio genuinely helps, where it slows things down, and what we need to change.”

The first weeks involved a natural onboarding curve: understanding the tech stack, getting familiar with Rebel AI Studio, and integrating AI tools into the workflow. Progress has picked up as the team has settled in.

While Codemate provides the industry context, like prioritisation, user-centred thinking, and practical decision-making, the students bring curiosity and fresh thinking that isn’t locked into old ways of working. Many already use AI tools in their studies to explore alternatives or validate their understanding; here, they get to apply those same habits in a real, shared codebase with proper version control and peer reviews.

AI in everyday work

The team uses AI throughout the workflow, from planning to testing and CI/CD. In planning, it helps break features into smaller tasks, draft user stories, and think through options. Implementation happens in a shared GitHub codebase where AI supports day-to-day coding and refactoring. In reviews and testing, it helps spot issues and speeds up writing tests and documentation, which then run through the normal automated CI/CD pipelines.

At the same time, the group is already confronting practical questions:

When is it faster to write code manually than to adjust a suggestion?
How do you ensure quality when tools can generate large amounts of code quickly?
What division of labour between people and AI actually works in a real sprint?

These questions surface in everyday decisions, and the team will return to them as the project progresses.

What Code Forward will follow

This first blog has stayed close to the starting point: who is involved, what the project is built on, and why the collaboration exists.

In the next parts of Code Forward, we’ll move closer to the day-to-day reality of the project and see how the first versions of the application take shape. Along the way, you’ll hear from both students and mentors — including moments that went smoothly and moments that challenged us.