Stefania Fresca

Stefania Fresca

Assistant Professor

Department of Mechanical Engineering, University of Washington

Biography

Stefania Fresca is Tenure-track Assistant Professor in Physics-based Machine Learning at the Department of Mechanical Engineering, University of Washington, Seattle, where she leads the FRESCA Lab (Scientific Machine Learning for Computational Engineering). Her research is supported by the National Science Foundation, UK Research and Innovation, and industry partners. She is also Visiting Assistant Professor at the Department of Computer Science and Technology, University of Cambridge. Previously, she was Junior Assistant Professor in Numerical Analysis at MOX (Laboratory for Modeling and Scientific Computing) - Department of Mathematics, Politecnico di Milano, Italy, within the Future Artificial Intelligence Research (FAIR) Project.
After carrying out her PhD in the framework of the ERC Advanced Grant Project iHEART (PI: Prof. Alfio Quarteroni) devoted to cardiac modeling, she spent two years as Post-Doctoral Research Fellow at MOX.
Her research interests and expertise include scientific machine learning, reduced order modeling, digital twins, and numerical approximation of PDEs, with several applications to engineering problems ranging from cardiac modeling (cardiac electrophysiology) to computational mechanics (fluid dynamics, flow control). Her current focus is on structure preserving neural networks, multi-scale deep learning, and deep reinforcement learning.

Interests
  • Scientific machine learning
  • Reduced order modeling
  • Surrogate-based optimization
Education
  • PhD in Mathematical Models and Methods in Engineering, 2021

    Politecnico di Milano

  • MSc in Mathematical Engineering - Computational Science and Engineering, 2017

    Politecnico di Milano - Université Pierre et Marie Curie (Sorbonne Universités)

  • BSc in Mathematical Engineering, 2014

    Politecnico di Milano

Experience

 
 
 
 
 
Department of Mechanical Engineering, University of Washington
Tenure-track Assistant Professor
September 2025 – Present Seattle, USA
 
 
 
 
 
Department of Computer Science, University of Cambridge
Visiting Faculty
Department of Computer Science, University of Cambridge
September 2023 – Present Cambridge, UK
hosted by Prof. Pietro Liò and Prof. Carola Schönlieb
 
 
 
 
 
Future Artificial Intelligence Research (FAIR) Foundation
Junior Assistant Professor
February 2023 – August 2025 Milano, Italy
 
 
 
 
 
Participation to the program “The mathematical and statistical foundation of future data-driven engineering”
 
 
 
 
 
Ernst & Young
Risk Advisory Intern
June 2017 – November 2017 Milano, Italy

Responsibilities include:

  • Design and data modelling of a Datamart
  • Data extraction activities through SQL
  • Automation of Data Quality processes through Access and VBA
 
 
 
 
 
Université Pierre et Marie Curie (Sorbonne Universités)
Exchange Program
September 2015 – March 2016 Paris, France

Publications

Quickly discover relevant content by filtering publications.
(2026). Audited surrogate gradients for efficient model-based flow control. Sim2Science: ML with Imperfect Scientific Models, 40th Conference on Neural Information Processing Systems (NeurIPS).

PDF

(2026). GraphEP: A calibrated data system and benchmark for learning cardiac electrophysiology. AI Data Readiness for Scientific Discovery (AIDaR), 40th Conference on Neural Information Processing Systems (NeurIPS).

(2026). Explainable deep learning-based classification of Wolff-Parkinson-White electrocardiographic signals. Frontiers in Physiology, 17, 1855555.

PDF DOI

(2026). Mesoscopic-informed residual reinforcement learning for adaptive CAV headway control in mixed-autonomy traffic. HAL preprint hal-05707752.

PDF Source Document

(2026). Graph surrogate modeling for closed-loop microscopic-macroscopic control of mixed traffic. HAL preprint hal-05676908.

PDF Source Document

Upcoming ­ Dates

Media & Communications

*
Interview @ DMAT
“Conosci chi fa ricerca” section on the Mathematics Department’s website at Politecnico di Milano.
Interview @ DMAT
Post in Coventor MEMS+ Blog
Using Machine Learning to Develop a Real-Time Model of a MEMS Disk Resonating Gyroscope.
Post in Coventor MEMS+ Blog
Seminar @ Machine Learning + X Seminars (CRUNCH Group, Brown University)
Deep learning-based reduced order models for parametrized PDEs.
Seminar @ Machine Learning + X Seminars (CRUNCH Group, Brown University)
Article in Enginsoft Newsletter - RESEARCH & INNOVATION
Deep learning-based reduced order models - the new frontier in numerical simulation for microsystems.
Article in Enginsoft Newsletter - RESEARCH & INNOVATION
Talk @ Mathematics of Deep Learning Workshop (Isaac Newton Institute, University of Cambridge)
POD-DL-ROM - a comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs.
Talk @ Mathematics of Deep Learning Workshop (Isaac Newton Institute, University of Cambridge)
Talk @ MCF2021 Congress
Deep learning-based reduced order models for the real-time approximation of nonlinear time-dependent parametrized PDEs.
Talk @ MCF2021 Congress
Talk @ 36th international CAE conference and exhibition
How medicine and engineering interrelate - a female bioengineering perspective.
Talk @ 36th international CAE conference and exhibition
Interview for iHEART project channel
How will artificial intelligence contribute to computational cardiac medicine of the future.
Interview for iHEART project channel
Interview ioDONNA magazine
A PhD student studies how to cure the heart with mathematics.
Interview ioDONNA magazine

Contacts

I am always happy to hear from students, researchers, and potential collaborators.
If you have a question, an idea, or a project in scientific machine learning, get in touch!