I work on flexible statistical modelling—P-splines, mixed models, computational statistics, flexible survival for recurrent events, and latent memory for longitudinal athlete monitoring—with applications in health and sports science. Below: main contributions with paper anchors, software, and a short trajectory.

→ People · Projects · Publications

P-splines, mixed models, and computational statistics

The methodological core of my work is flexible regression via penalized splines (P-splines) and their mixed-model representation, together with the computational statistics needed to make these methods usable at scale—multidimensional and adaptive smoothers, ANOVA-type interactions, GLAM-type array methods, and related estimation and prediction tools.

Selected anchors

• Fast smoothing parameter separation in multidimensional generalized P-splines (SAP)
• Out-of-sample prediction in multidimensional P-spline models
• Multidimensional adaptive penalised splines

→ Publications · S3M1P4R · AFTERAM

Flexible survival for recurrent events

A second methods line develops flexible survival and event-history models for recurrent events, with smooth effects, time-varying covariates, and exposure histories that matter in longitudinal monitoring settings.

Selected anchors

• Flexible modelling of time-varying exposures and recurrent events (JRSS-C)
• Prediction of sports injuries in football: a recurrent time-to-event approach

→ Publications · MATH4SPORTS · People · Software

Latent memory for athlete monitoring

An emerging line treats recent longitudinal history as a reusable statistical object—the Latent Memory Table—rather than only as a fixed lag or EWMA scalar. The aim is a finite-dimensional memory state that can be stored, queried, and reused across the modelling workflow, with SoccerMon as the empirical case study.

Selected anchors

• Learning Latent Memory States from Longitudinal Athlete Monitoring Data (arXiv:2608.06290) · arXiv

→ Publications

Growth curves, derivatives, and maturation

Applied work on longitudinal growth: curve and derivative estimation, maturation timing (including peak height velocity / APHV), and shape-constrained approaches for reference-to-individual prediction in youth sport and paediatric settings.

Selected anchors

• Derivative curve estimation in longitudinal studies using P-splines
• Bayesian hierarchical modelling of growth curve derivatives
• Peak height velocity and injury burden in circa-PHV soccer players

→ Publications · SPHERES · MATH4SPORTS

Patient-reported outcomes and joint survival

A long-running health line develops beta-binomial and related models for longitudinal patient-reported outcomes (PROs), including joint modelling with survival endpoints (e.g. COPD applications).

Selected anchors

• A beta-binomial mixed-effects model for longitudinal discrete and bounded outcomes
• Patient-reported outcomes and survival analysis of COPD patients (joint modelling)

→ Publications

Open tools are part of how the methodological work travels.

• injurytools — R toolkit for sports injury and illness data analysis (CRAN; GitHub). Developed with Lore Zumeta Olaskoaga; supports preparation, description, and modelling workflows for recurrent injury settings.
• PROregTMB — Template Model Builder implementations connected to beta-binomial / PRO modelling lines.

→ GitHub · Publications

Universidad Carlos III de Madrid (to 2011). Training and early research in statistics and mathematical engineering; PhD in Statistics (2010). First academic appointments as teaching assistant and lecturer.

CSIRO / Data61, Melbourne (2011–2014). Postdoctoral Fellow in mathematical and statistical sciences, with a strong applied-modelling orientation.

BCAM – Basque Center for Applied Mathematics (2014–2022). Led the Applied Statistics research line and, from 2020, coordinated Data Science and Artificial Intelligence in the Knowledge Transfer Unit—bridging academic methods with industry and public-sector problems (including SME advisory work and the Basque Modelling Task Force during COVID-19).

IE University (2023–). Tenure-track Assistant Professor in Statistical Sciences at the School of Science & Technology, continuing methods and applications in health and sports science, doctoral supervision, and competitive research programmes such as SPHERES.

Full appointment list → Resume · CV → PDF