PhD in Computer Science

In modern software industry, modelling plays a key role in the success of developing and maintaining complex systems and systems of systems. Appropriate software models can hide selective details of components and processes that may be difficult to observe or understand directly, which significantly facilitates engineers’ design, documentation and communication at higher levels of abstraction. However, nowadays even models themselves are becoming more and more complex due to the increasing complexity of the underlying systems being modelled. For example, beyond “classical” data and behavioural aspects, we now need to come up with new models to define the new types of UIs or all the smart features of a modern system. This hampers the adoption of model-driven processes as it reduces the return on investment of modelling activities due to the increase cost of modelling. This project aims to investigate Vibe Modelling to address the current challenges in model-driven software engineering. Here we define Vibe Modelling as a collaborative approach where humans leverage AI agents to design and refine complex charts, diagrams, and models. Instead of starting from a blank canvas or rigid templates, Vibe Modelling (ideally) enables users to express high-level goals, describe processes in natural language, or provide rough sketches, while letting AI assist in structuring, visualizing, and iterating on these ideas. Unlike automated code generation (or even vibe coding), Vibe Modelling is fundamentally about assistance and augmentation, allowing intuition and expertise to guide the AI at different engineering stages, increasing the visibility of software process, and improving the quality and maintainability of delivered software systems. Thus, this project expects to integrate the best of both domains (AI/LLMs and model-driven software engineering) to develop a theoretical framework and a practical methodology for implementing Vibe Modelling.