Matteo Cercola
PhD candidate at Politecnico di Milano · Visiting Researcher at Stanford ASL
About
I am a PhD candidate in Information Technology at Politecnico di Milano (Prof. Simone Formentin) and currently a Visiting Student Researcher at Stanford University's Autonomous Systems Lab (Prof. Marco Pavone), where I work on foundation models for autonomous driving.
My research sits at the intersection of machine learning and control: foundation models for autonomous driving and robotics, reinforcement learning from human feedback for LLMs, and diffusion-based generation for control design. I am broadly interested in making learning-based systems that reason well, respect human preferences, and come with guarantees.
News
- Apr 2026Diffusion-based controller synthesis via Youla parameterization accepted at IFAC 2026.
- Mar 2026Our paper on efficient RLHF via Bayesian preference inference is out in the IFAC Journal of Systems and Control.
- Jan 2026Started as a Visiting Student Researcher at the Autonomous Systems Lab, Stanford University.
- Jan 2026Regularized GLISp for sensor-guided human-in-the-loop optimization published in the IFAC Journal of Systems and Control.
- Jun 2025Presented our work on LLM-based traffic incident management at L4DC 2025, Ann Arbor.
Publications
Research Highlights
Foundation models for autonomous driving — Stanford ASL
Enhancing the reasoning of foundation models for decision-making and control in dynamic environments; fine-tuning 4B-parameter vision-language-action models on a 32× H100 cluster with PyTorch distributed training.
Agentic LLM decision support — MOVYON / Autostrade per l'Italia
An agentic decision-support system (LangChain, RAG) with optimization-backed guarantees, deployed for highway radio operators to accelerate incident response.
LiDAR segmentation — Indy Autonomous Challenge 2023
Real-time point-cloud segmentation (PyTorch + ROS2) deployed on a full-scale autonomous race vehicle, contributing to the team victory at the Monza Circuit.