Enea Monzio Compagnoni, MSc. MSc.

Enea Monzio Compagnoni, MSc. MSc.

PhD student at the University of Basel (stochastic optimization, deep learning). Previously in finance. I am persistent on hard math and relentless on experiments: theory-driven training improvements, RL for Robotics, and large-scale hyperparameter optimization.

I am looking for Research Internships (Summer 2026)!

If you are looking for a highly productive and motivated researcher, consider me:

  1. I do not give up when Math gets hard.
  2. I like finding unorthodox angles and validating them with clean experiments.

I enjoy turning theory into training improvements you can measure.

Presentation

I am a final-year PhD candidate in Stochastic Optimization for Deep Learning at the University of Basel (Switzerland), supervised by Prof. Dr. Aurelien Lucchi.

My main interests are stochastic optimization and deep learning, and more recently reinforcement learning for robotics.

I also have hands-on experience with VLMs/VLAs for robotics: at Flexion Robotics, I built an agentic framework that uses vision-language models to help robots execute multi-step tasks in real-world settings.

Lately, I have been playing around with:

  1. Stochastic optimizers for deep reinforcement learning, including large-scale HPO for PPO-style training and a promising modification to how AdamW is applied to the GRPO objective.
  2. New differentially private optimizers for LLMs, especially regimes where strong privacy is required.
  3. Reasoning efficiency in LLMs, including chain-of-thought compression.

Drop me an email :) eneamonziocompagnoni@gmail.com