Computational Mathematics · Geospatial Technologies
I turn messy spatial and temporal data into models across compression algorithms, clinical outcomes, and environmental risk, and build the tools to test them against reality.
Computational Mathematics graduate from Universitat Jaume I, currently completing an MSc in Geospatial Technologies through the Erasmus Mundus programme — a joint degree across NOVA IMS (Lisbon), the University of Münster (IFGI), and Universitat Jaume I.
My work moves between mathematical theory and applied implementation: I've optimised low-level memory algorithms for scientific data compression, analysed clinical datasets to flag hospital risk factors, and I'm now building a Bayesian spatio-temporal model of wildfire ignition risk across the Iberian Peninsula for my thesis, supervised by Jorge Mateu and Edzer Pebesma. I work comfortably in Python and R, and I like building tools other people can actually use — not just papers that describe them.
Joint degree spanning Portugal, Germany, and Spain. Thesis: a Bayesian spatio-temporal model of wildfire ignition risk on the Iberian Peninsula, built with INLA and benchmarked against a covariate-only baseline.
Investigated rescue failure patterns in hospital admissions — literature review, dataset analysis, and logistic regression / clustering to surface risk factors for patient deterioration.
Curricular internship improving matrix transposition and memory layout for the Blosc2 compression library, with benchmarks against NumPy. Results written up on the ironArray blog.
Read the write-up →An open-source R package for spatio-temporal event analysis, built from scratch around a vectorised Hawkes-process intensity kernel. Includes S3 classes for event sets and spatial grids, neighbourhood and counting utilities, two vignettes, bundled real-world datasets, and a full testthat suite with CI.