“Targeting this diversity across different use cases in a three-year project is really tough.”
— Xavier Franch, Work Package 5 Leader, HIVEMIND

In this video, Xavier Franch, professor at the Universitat Politècnica de Catalunya (UPC), shares how his team is driving Work Package 5, Smart System Specification, in the HIVEMIND project. He talks about their research group’s background, what they find most stimulating about the initiative, and what challenges they see ahead.

Watch the interview

Prefer reading?
You can find a full transcript of the interview below.

My name is Xavier Franch. I work at the Universitat Politècnica de Catalunya (UPC) in Barcelona, where I’m a full professor and the principal investigator of the Research Group on Software and Service Engineering.

We are a team of 10 to 12 principal researchers, along with a number of students working with us. We are leading Work Package 5, Smart System Specification, and we are also collaborating in other tasks across different work packages.

How does your expertise contribute to the objectives of HIVEMIND?
Our team has been working in software engineering for more than 20 years now. Recently, with the advent of artificial intelligence and machine learning in different areas, we’ve specialised in applying large language models and other types of networks to various stages of the software lifecycle, such as requirements elicitation, software architecture, design, and more.

What excites you most about being part of the project?
It’s a real challenge. There are two things I find particularly interesting about HIVEMIND. First, exploring the boundaries between humans and machines, to what extent we can simply delegate tasks to LLMs, or whether we still need to interact and contribute our own expertise. That’s still a matter of debate.
Second, there’s this holistic, end-to-end view, from requirements elicitation to final validation. Each phase involves different agents that must collaborate, and sharing these different perspectives will be a real challenge. But that’s also what excites us.

What impact do you expect HIVEMIND to have on industry and society?
From the industry point of view, we’re working with several use cases, each one quite challenging in its own domain. I expect every use case to have very particular needs, and if we’re able to address them successfully, the impact could be significant.
On the other hand, this also makes things more complex. Targeting this diversity across different use cases in a three-year project is really tough.
And from a societal perspective, I hope the project will also address ethical concerns, such as bias, gender issues, and similar challenges, which are among the main risks we face when using these kinds of technologies. I truly hope we’ll be able to take these aspects into account as the project progresses.