Claudio is a Research Scientist at the FutureTech Lab at MIT. His research sits at the intersection of mathematics and AI, building rigorous foundations for trustworthy and safe AI, from inference-time alignment and interpretability to fairness and accountability in machine learning models. He is particularly interested in safeguards and cybersecurity for AI agents, the economic and societal implications of deploying AI systems in the real world, and leveraging AI to accelerate discovery in science and mathematics. Claudio obtained his Ph.D. in mathematics and electrical engineering (summa cum laude) from the Technical University of Munich, where he worked with Felix Krahmer and Holger Boche. Before joining MIT, he was a postdoctoral researcher at Harvard's School of Engineering and Applied Sciences. He has also collaborated with lawyers and policymakers on AI governance, including contributing to the G20 Summit policy discussions. His research appears in leading venues in AI and machine learning, including NeurIPS, ICML, ICLR, CVPR, AISTATS and ACM FAccT.