I am currently a postdoctoral researcher at MIT, working with Stefanie Jegelka. Previously, I completed my PhD at EPFL, where I was supervised by Pierre Vandergheynst and worked closely with Andreas Loukas. I work at the intersection of combinatorial optimization and machine learning, where I focus on neural networks and gradient-based methods to solve challenging optimization problems over discrete structures.
Training neural networks with gradient descent to solve discrete optimization problems creates a bridge between beautiful ideas in discrete and continuous mathematics. At this interface, two directions particularly drive my research.
- Finding natural high-dimensional continuous formulations of discrete optimization problems that enable efficient gradient-based optimization.
- Developing data-driven algorithms for discrete optimization with guarantees on feasibility, efficiency, and approximation quality.
If you are a student interested in working with me on those topics, feel free to shoot me an email!
Outside of research, I love doing parkour and playing the drums. I will post more content related to my research on this website such as tutorials or practical guides for combinatorial ML. However, I may occassionally write something about other topics (love the Matrix Trilogy!).