Vito Antonio Pagone

Data Scientist – UBS

Mar 2024 – Present | Zürich, Switzerland

I work in the AI & Data Science team within Global Wealth Management in Zurich. Our small but mighty subgroup, Trading Analytics, collaborates closely with the Execution Hub, the place where client orders meet the market. My focus is on post-trade analytics, using machine learning to tackle real problems that make trading smarter and faster.

My daily toolkit revolves around Python and Airflow. I build and maintain data pipelines that keep our analytics flowing, Airflow is a brilliant ally for transforming internal data efficiently and transparently. I’m also part of the AI Journal Club, where we regularly discuss new research papers and explore how fresh ideas can shape our work.

Lately, I’ve been automating report generation for best execution, so our insights are always clear, consistent, and just a click away.

Machine Learning Intern – IBM Research

Nov 2023 – Jan 2024 | Zürich, Switzerland

Although a short experience, it was an incredibly inspiring one. I worked on the idea that if machine learning models are trained to predict physical phenomena from simulation data, they should also be given some prior knowledge of the underlying physics.

At IBM, I explored how atmospheric and climate equations could be embedded directly into the loss function of a forecasting network to guide convergence, especially during the early training epochs. It was fascinating to see how blending data and physics can produce more stable and interpretable results.

Beyond the technical side, I got to experience the IBM Research culture, collaborating with brilliant people from around the world. I left genuinely grateful for the opportunity and the mindset it gave me.

Machine Learning Researcher & Python Developer – ETH Zurich

Feb 2023 – Jan 2024 | Zürich, Switzerland

My time at ETH Zurich brought together teaching, coding, and research in a way that shaped how I think about science and communication. As a Research Assistant for the MSc course Introduction to Scientific Computing taught by Prof. Roman Vetter, I created interactive Jupyter notebooks to help students visualize algorithms and truly understand how they work, making complex numerical methods more tangible.

Later, as a Machine Learning Researcher, I continued my Master’s thesis work, developing Graph Physics-Informed Neural Networks (GPINNs) to improve field reconstruction accuracy in fluid dynamics. This phase focused on preparing the research for publication, an effort I pursued under the guidance of Dr. Bogdan Danciu. It was a deeply rewarding experience that combined curiosity, persistence, and the excitement of turning theory into something real.

Intern – MAN Energy Solutions

Sept 2022 – Mar 2023 | Zürich, Switzerland

This internship marked my first real step into the world of applied fluid dynamics and computational physics. I worked on a simulation framework to reproduce the supersonic two-phase flow of CO₂ through converging–diverging nozzles, a fascinating problem that taught me how physics and code can complement each other beautifully.

It was also my first time working in a company environment, and it helped me grow both technically and personally. I became much more confident with Linux systems and scientific programming, gaining hands-on experience in writing, testing, and running simulations at scale.