Claudio Fanconi

Ph.D. Student in Machine Learning, University of Cambridge

Hello! I am a third-year PhD student in Machine Learning and Artificial Intelligence at the University of Cambridge, advised by Mihaela van der Schaar.

Previously, I graduated with a Bachelor's and Master's from ETH Zurich in Information Technology and Electrical Engineering, with a focus on Machine Learning. I conducted my Master's thesis under the supervision of Tina Hernandez-Boussard at Stanford University, researching predictive uncertainty and natural language processing to identify patients at risk of acute care use.

In my PhD, I investigate how AI can enhance decision-making, focusing on:

Understanding and quantifying reasoning in LLMs (e.g. through inverse reinforcement learning).
Analysing frameworks for collaboration between multiple LLMs and experts.
Developing methods for personalisation to improve LLM effectiveness.

Reach out to me with anything you want to talk about.

Additionally, please check out the Fanconi Cancer Foundation and their initiatives against Fanconi Anemia.

Updates

2026/03
First-author paper accepted (oral) at the ICLR Workshop on LLM Reasoning (paper): adversarial inverse RL for eliciting and quantifying reasoning in LLMs.
2026/03
Paper with Paulius accepted (spotlight) at the ICLR Workshop on Recursive Self-Improvement (paper): tiny autoregressive recursive networks.
2025/09
First-author paper accepted (poster) at NeurIPS 2025 (paper): cost-efficient cascaded LLM systems for decision-making.
2025/06
Co-first-author paper accepted (oral, top 10%) at the ICML 2025 Workshop on AI Alignment, together with Kasia (paper): few-shot steerable alignment of rewards and policies with neural processes.
2024/09
Co-first-author paper accepted (poster) at NeurIPS 2024 (paper): LLMs generating optimisation algorithms for preference learning.
2024/04
Started PhD in Machine Learning at the University of Cambridge under the supervision of Mihaela van der Schaar.

At A Glance

Current Role Ph.D. Student in Machine Learning
Institution University of Cambridge
Advisor Mihaela van der Schaar
Main Areas
- Reasoning in LLMs
- Human-AI Decision-Making
- Inverse Reinforcement Learning
Previous Industry Role Research Scientist Intern, Sony AI