Hello! 👋

I am Matteo, a researcher in machine learning.

My research focuses on reinforcement learning applied to multi-agent and multi-robot systems.

I am currently working on training LLM agents for long-horizon tasks using reinforcement learning at Meta.

I obtained my PhD from the Prorok Lab at the University of Cambridge, were I studied heterogeneity in multi-agent and multi-robot systems and developed the VMAS simulator. For this research, I employed techniques from the fields of Multi-Agent Reinforcement Learning and Graph Neural Networks. During my PhD, I joined PyTorch at Meta, where I created BenchMARL and helped develop TorchRL.

For my master, I investigated the problem of transport network design for multi-agent routing.

See my CV and the (more verbose) academic CV.