Juan is currently a PhD student at the Fluid Dynamic Technologies group of the University of Zaragoza in Spain under the supervision of Dr. Javier Murillo and Prof. Pilar Garcia-Navarro.

Juan's research focuses on modeling 1-D networks of flows, with application to both hydrodynamics and haemodynamics. His previous work has been focused on building robust finite-volume methods based on the exact solution of the Riemann problem at junctions and external boundaries to tackle transitory scenarios. This research has direct applications in river basin modelling and flood forecasting, as well as prediction of the haemodynamic response of postural changes.

Recently, he has started to apply Physics-Informed Machine Learning techniques to solve the inverse problem of parametrisation needed to construct personalised models. During this work, he has become interested in the potential of Physics-Informed Machine Learning to combine partially-known physical models and some limited data to overcome the uncertainty in the parameters and the scarcity in the measurements.

Juan has visited the HaeMod group in London to work closely with the team on the applications of Physics-Informed Neural Networks for arterial signal recuperation. During this time, he has worked on methods to recover central signals and arterial parameters from the least possible amount of data. For this, he has used reduced Windkessel models of arterial trees, which, due to their simplicity, allow for non-expensive training of the networks while maintaining sufficient accuracy.