Profile

I’m a physics researcher in experimental particle physics working on the CMS collaboration at the LHC, CERN, with a primary focus on applying deep learning (artificial intelligence (AI)) to a variety of problems that are unique to our field of study. Currently I am based at the University of Padua, in Northern Italy, and have been living here for just over a year.

I recently completed my PhD in physics in Lisbon, Portugal during which I worked as part of a European-funded international collaboration of 10 fellow PhD students (AMVA4NewPhysics). Together we studied the application of AI to many different tasks at the LHC. The programme included a lot of travel (pre-COVID, luckily), and extended stays at various universities and science facilities (including 3-months in Moscow, Russia during the winter). I also blog on my personal website, and previously wrote for the AMVA4NewPhysics site.

Outside of work, I play electric guitar and bass (mostly metal and rock), occasionally record and mix music, cook, and listen to music. I also practice Shorinji Kempo, a Japanese martial art, and currently hold the rank of first-level black-belt, having trained in England, Scotland, Portugal, and Japan.

I work on a range of activities, but they all revolve around developing and applying state-of-the-art computer algorithms to the study of the universe at its most fundamental levels (the interactions between particles at distances smaller than atoms).

These computer algorithms aim to allow the computer to learn how to solve challenging problems by itself; we just show it enough examples of what to do, and eventually it becomes able to perform better than we can. This is referred to as, machine learning or deep learning, and in the media as “Artificial Intelligence (AI)”

Within the CMS (Compact Muon Solenoid) Collaboration at CERN (The European Centre for Nuclear Physics research), we use this technology to assist us in identifying very rare interaction processes, which can allow us to test our theories of how the universe behaves. To help my colleagues to apply such algorithms in their own work, I develop software packages to provide easy access to advanced techniques.

Within the MODE (Machine-learning Optimized Design of Experiments) Collaboration, I help to develop software frameworks which allow designers of experiments to specify their end goal, and then have the compute automatically optimise the experiment to provide the best possible configuration.

I’ll aim to wake up around 7:30 and after shower, breakfast, and coffee, begin work at 9:00 from my home office.

This starts with checking my “To Do” lists and calendar to plan out what I need to achieve during the day. Activities could be: writing code for my analysis work or the software packages I develop; writing up my results for publication in scientific papers; preparing presentations and tutorials for upcoming events; following online talks; reading and reviewing papers, and trying out new analysis techniques; and also meeting with colleagues via video chat to discuss progress and plans or to code together. Generally I have a lot of freedom in what I work on during the day, provided the important stuff gets done, and the other activities are beneficial somehow. Pre-COVID I would often be travelling to work with colleagues or to attend conferences and workshops.

I work until 17:00, taking 10 minutes off for lunch, and then aim to disconnect from “work mode” through my hobbies.