The University of L’Aquila (UDA) is a public teaching and research institution established in 1952 and located in L’Aquila, in central Italy. The university offers a wide range of academic programs across disciplines such as engineering, medicine, sciences, humanities, and economics. It is committed to academic excellence, fostering cutting-edge research, and promoting international collaboration through partnerships with institutions worldwide.
Within the Department of Information Engineering, Computer Science and Mathematics (DISIM), the Software Engineering (SWEN) group is at the forefront of research on a wide array of software engineering challenges. SWEN group conducts interdisciplinary research combining software engineering, artificial intelligence, and software architecture. Their work focuses on both functional and non-functional aspects of software systems, with applications in quality assurance, recommender systems, model-driven engineering, and empirical SE.
In the MOSAICO project, UDA brings this strong expertise to build core infrastructure for managing AI agents across the software development lifecycle.
What is UDA working on in MOSAICO?
UDA leads Work Package 2, which is dedicated to designing and implementing the Repository of AI Agents for Software Engineering, a central component of the MOSAICO platform.
This repository is more than just a catalog: it is a smart, structured, and quality-aware environment that supports the discovery, evaluation, and reuse of AI agents tailored for SE tasks. Through the repository, MOSAICO aims to make the growing ecosystem of LLM-based tools more navigable and trustworthy for developers.
The repository will feature:
- Comprehensive Metadata: Each AI agent is annotated with detailed descriptions of its capabilities, training data, limitations, supported file types, and ideal use cases.
- Advanced Search and Filtering: Developers can filter agents based on performance indicators, supported technologies, and specific SE task suitability.
- Quality Assessment Models: UDA is developing customizable, automated evaluation models to measure agents’ effectiveness through KPIs like accuracy, latency, and fairness.
- Benchmarking Framework: The repository will include curated benchmarks to evaluate and compare AI agents consistently across tasks.
- Recommendation System: Intelligent agent suggestions will be based on past usage patterns, performance data, and developer preferences.
With this work, UDA ensures that AI agents in MOSAICO are discoverable, assessable, and ready for real-world deployment in complex SE workflows.
Meet the face behind the project
Davide Di Ruscio is Full Professor at the University of L’Aquila and member of the SWEN group and of the Model-Driven Engineering and Software Mining Laboratory. His research interests span model-driven engineering, recommender systems, AI-based development environments, and automated software analysis.
He is supported by a dynamic team of researchers who have developed foundational tools like Jjodel (a novel metamodeling platform), MDEForge (a repository of modeling artifacts enriched with metadata and machine learning-based quality evaluations), and several recommenders from software engineering.
In MOSAICO, Prof. Di Ruscio leads the design and implementation of the AI Agent Repository, contributing deep knowledge from past EU projects and national initiatives focused on quality-aware software reuse and AI-based development assistance. Key initiatives from the work done include contributions to projects like CROSSMINER, EMELIOT, TRex-SE, and FRINGE, showcasing their strength in repository mining, trustable recommendation systems, and the use of machine learning for fairness and effectiveness in software tools.
About UDA
The University of L’Aquila is a prominent academic institution known for its high-quality education and impactful research. With over 20,000 students and 600+ faculty members, it maintains a strong focus on innovation, interdisciplinary collaboration, and international partnerships. As part of the MOSAICO consortium, UDA plays a central role in enabling the discoverability, assessment, and integration of AI agents for software engineering.
