Bio

I work across AI research, statistical physics, neuroscience, and scientific software. Over the years this has meant moving between incomplete data, mathematical models, research code, institutional constraints, and people who do not naturally speak the same language: physicists, neuroscientists, software developers, finance stakeholders, experimental researchers.

I tend to work where an idea has to survive contact with data, code, collaborators, and deadlines. A model is only part of that work. The assumptions have to become visible enough for a group to trust the result, question it, or decide that it should not be used.

Across nearly two decades I have also learned by proximity: in research groups, summer schools, workshops, and long conversations with unusually sharp people in physics, neuroscience, psychiatry, and AI. Some of my best work has come from carrying fragments across rooms: a formal model, an experimental constraint, a clinical intuition, a piece of code, a mathematical analogy that becomes useful years later.

  • Fellow of the ThinkingAboutThinking Society, an Oxford-based group working at the intersection of mathematics, neuroscience, and AI.
  • 2022-today Senior AI Researcher at Ipazia SpA, Milan, Italy. Research and engineering on modern AI systems, with a focus on translating mathematically grounded ideas into methods and software that can stand up to use. The role moves between hypothesis formation, model evaluation, scientific communication, and day-to-day decisions under uncertainty.
  • 2020-2022 Senior analyst/data scientist at Prometeia, Milan, Italy. Data science, distributed computing (Spark), anti-money laundering. Worked in large cross-functional teams on finance projects, aligning technical work with stakeholder needs, risk constraints, and operational delivery.
  • 2019 PostDoc in Statistical Physics for complex networks analysis in neuroscience.
  • 2017 Phd in Applied Physics (Nanoscience and advanced technologies), University of Verona, Italy. Thesis title: “Community detection in the modular structure of brain functional connectivity networks”.
  • 2008 M.Sc in Biomedical Physics, Grade 110/110 cum laude, University of Trento, Italy.
  • 2003 B.Sc in Physics, University of Trento, Italy.

Seminars and conferences (main)

  • November 2024 ICAIF conference on AI in finance, Brooklyn, NY (USA).
  • July 2019 Lipari Workshop on Complex Systems: Scale resolved analysis of complex networks Download
  • June 2018 Netsci2018, Paris. Thermodynamics of network model fitting with spectral entropies. Download
  • May 2016 CCS2016 Amsterdam, Community detection in brain functional networks beyond the resolution limit. Download

Experience

  • Senior Analyst Data-scientist and data-engineer at Prometeia, Milan. Worked in large collaborative teams, shoulder to shoulder with stakeholders over finance-related projects from NPL prediction to anti-money laundering methods rooted in rigorous data science algorithms. Contributed where modeling met data engineering and delivery, helping translate complex analytical work into solutions usable by business and risk teams. The work forced a practical negotiation between model quality, interpretability, project constraints, and institutional risk.

  • Now (2017-2019): Postdoctoral fellow

Developed new computational methods for the analysis of modular structure of brain functional connectivity networks with the tools of quantum information, information theory and complex networks theory. Research done at the BrainetLab of Angelo Bifone, in the CNCS, Istituto Italiano di Tecnologia. Developer of the networkqit package, a Python package to do computations in the framework of spectral entropies of complex networks. This work combined methodological research with scientific software development and day-to-day collaboration with experimental researchers. It also trained a particular discipline: keeping mathematical assumptions visible while adapting methods to empirical data.

  • Spring-Summer 2017

Visiting researcher at Lorentz Institute for Theoretical Physics, Leiden, The Netherlands. Working on random matrix theory, generative models of human brain connectivity and community detection, under the guidance of Prof. Diego Garlaschelli.

  • 2012-2017 System administrator for high performance computing

Maintenance of a small computational cluster, experience in high performance computing, parallel computing (Matlab parallel computing toolbox) and CUDA parallel capabilities. Responsible for keeping shared computational infrastructure reliable and available for research work, balancing individual research needs with shared operational constraints.

  • 2011 Software consultant at Brown University (USA)

Development, testing and installation of a full virtual reality system at the cognitive neuroscience department of Brown University, Providence (RI), USA http://bit.ly/IoyXV2. Delivered an end-to-end technical system in an international research setting, from implementation to testing and installation. The project required local autonomy, remote coordination, and a clean handover to researchers who would use the system after delivery.

  • 2010–2017 Scientific programmer at Istituto Italiano di Tecnologia

Working as the main developer of the C++ and Matlab/Psychtoolbox codebase for cognitive neuroscience experiments at the CNCS@IIT research group. Expertise in big C++ projects development and maintenance as well as robust foundations in algorithms, numerical analysis, matrix computation and data structures. Acted as a technical reference point for shared experimental software, helping keep research code maintainable and useful across projects and collaborators. Much of the role was quiet leadership: stabilizing tools, clarifying requirements, and keeping scientific work from being blocked by fragile code.

  • 2008-2010 Research fellow on Machine learning and intelligent optimization

Working on human activity recognition with statistical pattern recognition tools such as SVMs, ANNs, and probabilistic graphical models (HMMs and LCRFs). Experience in machine learning and numerical optimization http://bit.ly/IokSWd. This was my first sustained experience working on machine learning systems with both theoretical and applied constraints in mind.