| layout | page |
|---|---|
| title | About |
| permalink | /about/ |
| description | Background, interests, and current work in uncrewed aerial systems, robotics, and AI. |
| body_class | dark-page |
I work on autonomous machines, and on how intelligent behavior gets designed and then made to work outside a lab.
Today I am Head of Software at 24 Industries, a startup that builds uncrewed aerial systems. The work covers drone software and Ground Control Stations: architecture, autonomy, simulation, software quality assurance, and compliance with aerospace standards. It also covers the parts nobody puts in a talk: making behavior reliable, fitting software to hardware limits, hiring, and keeping a growing team able to follow its own code.
Before that I spent years in logistics robotics, where I led a team of more than 20 engineers responsible for robot architecture design, robot behavior development, simulation, and software and hardware quality assurance. I have worked mainly in Python, C++, and Prolog, on ROS.
I came to robotics through physics and then artificial intelligence. During my PhD in Robotics and Artificial Intelligence at Sapienza University of Rome, and a visiting period at Carnegie Mellon University, I worked on semantic mapping, human-robot interaction, task acquisition, and knowledge representation.
During my studies I developed a deep interest in artificial intelligence and robotics, and above all in the cognitive side of the subject. One question has fascinated me since: how can a machine, a single robot or a multi-robot system, be given intelligent behavior that lets it learn and reason about how to act in a complex world? That question holds others that humankind needs, at least, to attempt to answer. What do we mean by the adjective intelligent? What gives a specific entity that attribute? Will we ever be able to instill that trait in a machine? I would like to spend my life contributing to the answers.
That interest pulls me toward systems that keep working after deployment, which is a harder test than a result on paper. Over the years it has taken the form of research papers, patents, European projects, open-source contributions, and product engineering. This site collects those projects and publications, and some of the ideas behind them.