I bridge process mining, machine learning, and explainable AI to enable transparent what-if and prescriptive analytics, with applications in healthcare and industry.
Research Fellow at the University of Padua, Department of Mathematics "Tullio Levi-Civita".
I'm Francesco, an AI and process science researcher at the University of Padua. My work bridges process mining, machine learning, and explainable AI to enable transparent what-if and prescriptive analytics in real-world settings. I focus on data-driven process simulation models for operational decision-making, with applications in healthcare and industry, and develop open-source tools to translate research into practice.
Extracting insights from event logs to discover, monitor, and improve processes.
Predictive and generative models applied to process analytics and real-world challenges.
Transparent, explainable models for trustworthy decision support.
Formal frameworks, probabilistic models, and mathematical foundations.
End-to-end pipelines from data acquisition to actionable insights.
Data-driven discrete-event simulation for what-if analysis.
Python, SQL, R, JavaScript, C, C++
PyTorch, TensorFlow, scikit-learn, XGBoost, SHAP, River
Docker, Git, GitHub Actions, Flask, FastAPI, AWS
LangChain, prompt engineering, Claude Code, agentic workflows
PostgreSQL, MongoDB, REST APIs, XES/OCEL event logs
Matplotlib, Plotly, pandas, Jupyter
University of Padua
University of Padua
University of Padua
University of Padua
Work experience, education, and achievements.
Research on the discovery and enhancement of process simulation models using explainable machine learning techniques, generative and agentic AI, with applications to the optimization of healthcare processes, particularly in emergency medicine.
Research in process mining, process simulation, and data-driven modeling, with a focus on the discovery, repair, and improvement of simulation models using event data. Leveraged data science and AI techniques for online model adaptation and process optimization, with applications to business and healthcare processes.
Visited the Process and Data Science (PADS) group, conducting research on white-box process simulation models. Developed ProSiT, a tool for configurable process simulation. Supervised by Prof. Wil van der Aalst.
Visited the Fraunhofer Process Mining Group in Aachen, conducting research on white-box simulation models and their adaptation in online settings. Supervised by Prof. Wil van der Aalst and Dr. Gyunam Park.
Teaching Assistant for the Database course in the B.Sc. in Computer Science, supporting students during laboratory sessions on PostgreSQL and pgAdmin.
Developed a framework combining supervised and unsupervised machine and deep learning methods for video anomaly detection to identify sticking events in the continuous casting process of steel manufacturing.
Curricula: Computer Science for Societal Challenges and Innovation
Thesis: Towards Reliable and Explainable Process Simulation Models: Discovery, Refinement, and Applications
Supervisor: Prof. Massimiliano de Leoni — Cosupervisor: Dr. Silvia Gabrielli
Thesis Reviewers: Prof. Marlon Dumas, Prof. Andrea Marrella
Thesis: Semi-supervised Deep Learning methods for Video Anomaly Detection applied to sticking identification during steelmaking continuous casting process
Supervisor: Prof. Michele Rossi
Thesis: Enlargement of filtrations in discrete time and applications to finance
Supervisor: Prof. Claudio Fontana
Speaker: Reliable and Configurable Process Simulations via Probabilistic White-Box Models (main conference)
Speaker: ProSiT: A Tool for Interactive and Transparent Process Simulations (demos & resources)
Speaker: Online Discovery of Simulation Models for Evolving Business Processes (main conference)
Speaker: Healthcare Process Optimization via Simulations: An Emergency Department Case Study (poster)
Attendee
Speaker: Repairing Process Models Through Simulation and Explainable AI (main conference)
Organizer
Speaker: Investigating the Influence of Data-Aware Process States on Activity Probabilities in Simulation Models (main conference)
Attendee
Research on the discovery and enhancement of process simulation models using explainable machine learning techniques, generative and agentic AI, with applications to the optimization of healthcare processes, particularly in emergency medicine.
Research in process mining, process simulation, and data-driven modeling, with a focus on the discovery, repair, and improvement of simulation models using event data. Leveraged data science and AI techniques for online model adaptation and process optimization, with applications to business and healthcare processes.
Visited the Process and Data Science (PADS) group, conducting research on white-box process simulation models. Developed ProSiT, a tool for configurable process simulation. Supervised by Prof. Wil van der Aalst.
Visited the Fraunhofer Process Mining Group in Aachen, conducting research on white-box simulation models and their adaptation in online settings. Supervised by Prof. Wil van der Aalst and Dr. Gyunam Park.
Teaching Assistant for the Database course in the B.Sc. in Computer Science, supporting students during laboratory sessions on PostgreSQL and pgAdmin.
Developed a framework combining supervised and unsupervised machine and deep learning methods for video anomaly detection to identify sticking events in the continuous casting process of steel manufacturing.
Curricula: Computer Science for Societal Challenges and Innovation
Thesis: Towards Reliable and Explainable Process Simulation Models: Discovery, Refinement, and Applications
Supervisor: Prof. Massimiliano de Leoni — Cosupervisor: Dr. Silvia Gabrielli
Thesis Reviewers: Prof. Marlon Dumas, Prof. Andrea Marrella
Thesis: Semi-supervised Deep Learning methods for Video Anomaly Detection applied to sticking identification during steelmaking continuous casting process
Supervisor: Prof. Michele Rossi
Thesis: Enlargement of filtrations in discrete time and applications to finance
Supervisor: Prof. Claudio Fontana
Finalist
Premio Promotori Winner
Fondazione Italiana Accenture — Sustainability & Innovation
Tech conference — Startups, AI & Innovation
Peer-reviewed papers in process mining, AI, and simulation.
Open-source tools to translate research into practice.
Full-stack web application for data-driven business process simulation. Discovers simulation parameters from event logs using explainable machine learning, and enables interactive what-if scenario analysis with visual dashboards.
Python framework for business process simulation based on Petri Nets. Automatically learns simulation parameters from event logs, including transition probabilities, resource allocation, and working calendars.
Repair framework for Petri Net process models combining simulation and Explainable AI. Identifies structural differences between real and simulated traces to pinpoint and correct model inaccuracies.
Python framework for estimating activity start timestamps in event logs where only completion times are recorded. Includes evaluation benchmarks.
Stochastic Petri Net simulator. Reads PNML models and event logs, learns stochastic firing patterns using machine learning, and generates synthetic execution traces conforming to observed process behavior.
This personal website. Built with vanilla HTML, CSS, and JavaScript.
Interested in my research, projects, or a collaboration?
Find me here: