Michele Garibbo

Ciao!

I am a Deep Learning Scientist in AI for Protein Design led by Noelia Ferruz at the Centre for Genomic Regulation. Previously, I was a Postdoctoral Research Associate in Neuroscience and Machine Learning in the Neural & Machine Learning Group at the University of Oxford. I completed my PhD in Machine Learning and Neuroscience at the University of Bristol under the supervision of Laurence Aitchison (primary), Casimir Ludwig, and Nathan Lepora.

---

Research interests

My research focuses on end-to-end training of autoregressive protein language models, with particular emphasis on large-scale pretraining and reinforcement learning–based alignment. I also explore genomic foundation models (e.g., Evo2) and their applications to de novo design of functional elements such as piggyBac transposons.

Previously, I worked on reinforcement learning applications to understanding motor learning in neuroscience, as well as deep RL research aimed at improving value estimation in Actor–Critic algorithms.

More broadly, I am interested in how machine learning and computational models can provide principled frameworks for understanding and designing biological systems.