● PHD STUDENT · AIMAGELAB, UNIMORE
I research how AI connects vision, language and knowledge. Outside research, I enjoy teaching, sharing what I know, and playing basketball.
01 / ABOUT
I am a PhD student in Information and Communication Technologies at AImageLab, University of Modena and Reggio Emilia.
My work explores multimodal large language models and retrieval-augmented generation: how models find useful evidence, reason over it, and answer questions about the visual world. My research interests also include video understanding and retrieval-augmented VideoQA.
I joined the PhD programme in November 2025, after graduating in Artificial Intelligence Engineering at UNIMORE. I also spent three years teaching computer science at Liceo Manfredo Fanti in Carpi, the high school I graduated from.
UNIMORE · AImageLab
110/110 with honours · UNIMORE
110/110 with honours · UNIMORE
Previously at E4 Computer Engineering, I worked on advanced RAG systems for on-premise generative AI, connecting my research interests with real-world enterprise data.
02 / PUBLICATIONS & PREPRINTS
Answering visual questions sometimes takes more than looking at an image. ReAG retrieves external knowledge, filters distracting passages, and uses reinforcement learning to improve reasoning over the remaining evidence.
Better answers on Encyclopedic-VQA and InfoSeek, grounded in retrieved knowledge.
* Equal contribution.
What if a model could take a second look at the evidence that matters? Look Twice uses a model’s own attention to highlight relevant image regions and text, remove distractions, and guide its answer without additional training.
Evidence selection at inference time, without parameter updates or architectural changes.
03 / RESEARCH INTERESTS
How can AI use the right evidence to better understand images, videos and the questions we ask?
Connecting vision and language to build models that reason across different sources of information.
VISION + LANGUAGEFinding, filtering and integrating external knowledge to make generated answers more informed and grounded.
RETRIEVAL + REASONINGAnswering questions about images when the answer requires knowledge beyond what is directly visible.
IMAGES + KNOWLEDGEUnderstanding events over time, with a particular interest in applying RAG to video question answering.
VIDEO + RAG04 / TEACHING
I enjoy helping others make sense of something new, and learning from their questions along the way.
Teaching has been a rewarding part of my life. I like finding a clear way to explain an idea and seeing it start to make sense to someone else.
I taught computer science at Liceo Scientifico Manfredo Fanti in Carpi for three years, from 2023 to 2026. Returning to the school I graduated from, and sharing some AI knowledge with fellow teachers along the way, made the experience especially meaningful.
I was responsible for three to four classes, from planning lessons, homework and tests to grading, parent meetings and class councils.
05 / BEYOND RESEARCH
Basketball is a big part of my life outside research and teaching. My experience includes competing in Italy’s Serie C with Pallacanestro Scandiano.
The game keeps me grounded in teamwork, discipline and the habit of showing up, learning, and trying again.
My basketball profile ↗