Mattia Scardecchia, from Grottaferrata to New York
When you think of a talent in artificial intelligence, it’s tempting to picture the classic “nerd” glued to a screen and a bit awkward in real life. Mattia Scardecchia, however, immediately dispels that stereotype. He’s cheerful, athletic, and curious about everything. And he has a rare quality: he can grasp the essence of complex concepts and make them simple, without trivializing them. A skill that, according to those who know him, he already possessed as a child, when he was “the one who was good at explaining things to others.”
Today, Mattia is in New York, in his first year of doctoral studies at NYU’s Courant Institute. Along the way, he attended Bocconi University, worked at Sony AI in Zurich, and authored a paper featured on the cover of Nature that attracted a great deal of attention: the one on Ace, the autonomous robot capable of defeating top-level table tennis players.
But the interesting thing about his story is that it all seems to have begun long before the boom in generative AI and the global race toward artificial intelligence. It stems from an almost aesthetic passion for mathematics.
Mathematics as a language for understanding the world
Mattia says he was drawn to numbers from a very young age, when he used to plan to count all the way to infinity. What struck him was the beauty of mathematics: an inexhaustible web of hidden patterns waiting to be explored. Over time, he also began to appreciate its versatility: a language for describing reality, understanding its mechanisms, and influencing it. It was this combination of abstraction and concreteness that led him, after high school, to leave Grottaferrata, not far from Rome, to pursue the bachelor’s degree program in Artificial Intelligence at Bocconi University, which was then in its infancy.
What convinced him was not only the subject of AI itself, but the way the program was structured: a strong mathematical foundation, an interdisciplinary approach, and a rare balance between theory and practical applications. It was a path that seemed tailor-made for his learning style, always driven by curiosity rather than an obsession with exams.
As early as his second year, a major turning point occurred: his collaboration with Professor Riccardo Zecchina, one of Italy’s leading scholars in the fields of statistical physics and artificial intelligence. It was there that Mattia truly entered the world of research, learning not only the technical tools but also the scientific method: observing a problem, simplifying it, and seeking its hidden structure.
His master’s degree in AI thus represents a natural continuation of the path he began during his undergraduate studies. These were two years of intense growth, divided between university research, work at Bending Spoons, and studies at Bocconi and TUM. And it was precisely during this period that the opportunity arose that would change the trajectory of his career.
The robot that plays table tennis and the challenge of physical intelligence
During his master’s program, Mattia joined the Zurich-based Sony AI team, which was working on Ace, the robot featured in the study published in Nature. This project tackles one of the most difficult challenges in contemporary artificial intelligence: bringing AI from the virtual world into the physical world, which is often dynamic and unpredictable. Table tennis is a perfect testing ground for this challenge: at high levels, the ball can travel at over 20 meters per second and spin at more than 160 revolutions per second; every shot requires immediate perception, prediction, and reaction, with extreme precision. As the authors of the paper explain:
“Real-time physical sports, such as table tennis […], continue to represent a major challenge yet to be addressed, due to the rapid, precise, and competitive interactions they require.”
Ace combines computer vision, reinforcement learning, and advanced robotics to take on professional players under official conditions. And the results are surprising: the robot manages to defeat several elite players, demonstrating a capacity for adaptation and coordination that seemed unrealistic just a few years ago.
For Mattia, who contributed to the development of the robot’s control system using reinforcement learning techniques, working on this project means operating at the intersection of mathematical theory and practical application. It’s not simply about programming a robot, but about building systems capable of interpreting and making decisions in the real world, within dynamic and unpredictable environments.
From Zurich to New York: The future lies in “World Models”
Today, Mattia is at New York University, where he studies so-called “world models,” one of the most promising areas of current AI research. The idea is to develop systems capable of constructing an internal representation of the world: models that allow artificial intelligence to predict what will happen after an action, imagine alternative scenarios, and plan future behaviors. In practice, this is a step toward AI that is less “reactive” and more similar to humans in the way it understands its surroundings. This line of research could have applications everywhere, from autonomous robotics to smart vehicles, all the way to next-generation virtual assistants.
At the same time, Mattia continues to collaborate with researchers at Bocconi on a line of research at the intersection of artificial intelligence, statistical physics, and computational neuroscience. The goal is to study learning models that more closely mimic the brain, which can offer both tools for better understanding biological systems and energy-efficient AI algorithms on dedicated hardware.
His goal, he says, is to continue conducting research on artificial intelligence, whether at universities or in cutting-edge technology companies. But beyond the results he has already achieved, what is most striking is the way in which he approaches this journey: with an open, interdisciplinary mindset, far removed from the idea of study as the mere accumulation of knowledge.
Perhaps it is precisely this combination—mathematical rigor, curiosity, and the ability to get to the heart of problems—that has taken him from the classrooms of Bocconi to the laboratories of New York, via one of the world’s most advanced robotics projects.