Pomezia, Latium, Italy
I am a Ph.D. student in Data Science at Sapienza University of Rome, specializing in computer vision, embodied AI, and 3D human motion analysis. My work focuses on human-centered AI, robotic manipulation, motion synthesis, and multi-agent systems. I have experience in both academic research and startups, including collaborations with TU Darmstadt and Babelscape. My contributions resulted in different publications at top venues. Beyond research, I am passionate about AI that has a practical impact on human daily life, with the goal of making technology more accessible and meaningful to society.
Supervisor: Fabio Galasso. Research interests: egocentric vision, procedural learning, motion synthesis, multi-LLM-agent systems, computer vision.
Teaching assistant for the Advanced Machine Learning course, taught for the M.Sc. in Data Science. Implemented and evaluated the homeworks for the course.
Research in the fields of egocentric vision, multi-agent systems, deep generative models, and metanetworks. Mainly focused on extracting task graphs from egocentric cooking scenes.
Worked on in-context learning for bimanual robotic manipulation, leading to BiCICLe, a multi-agent leader-follower framework for training-free coordinated control. Supervised by Georgia Chalvatzaki and Fabio Galasso.
- Recognising dialect expressions used in different countries that speak the same language, both for English and Portuguese. Achieved 87% accuracy with a Java model that compares BabelNet’s senses starting from Wiktionary definitions. - Labeling BabelNet’s synsets based on their closeness to some main categorical synsets.