Tomas Guija-Valiente
PhD Student @ PROMISE–LAIMBIO, Universidad Rey Juan Carlos. Machine Learning, Computer Vision and Digital Twins. Open-Source Contributor.
Hi, and welcome to my personal website! My name is Tomás, and I am a PhD student at PROMISE–LAIMBIO, Universidad Rey Juan Carlos. My background lies at the intersection of mathematics and computer science, which I studied at Universidad de Sevilla, before specializing in artificial intelligence through an MSc at Universidad Politécnica de Madrid.
My current research explores machine learning, computer vision and generative AI for longitudinal modeling, patient simulation and digital twins. I am particularly interested in developing generalizable methods that extend beyond medical imaging: I believe the field should not only benefit from advances in machine learning and computer vision, but also help drive them.
Alongside my research, I actively contribute to open-source scientific software. Through Google Summer of Code 2026 with the Python Software Foundation, I contributed registration methods, performance improvements, bug fixes, tests and documentation to DIPY, an open-source Python library for diffusion MRI and medical image analysis. During the project, I was mentored by Serge Koudoro, Atharva Shah and Jong Sung Park.
Since then, I have expanded my open-source collaborations through merged contributions to Kornia, Hugging Face PEFT, timm and NVIDIA MedTech, with ongoing work in MONAI.
Latest news
| September 2026 | Presented two papers at MICCAI 2026: RARF, which introduces region-aware rectified flows for 3D brain MRI inpainting, and SynthRCT, a framework for generating synthetic repeat CTs through conditional deformation modeling. |
|---|---|
| September 2026 | Began a PhD at PROMISE–LAIMBIO, Universidad Rey Juan Carlos, researching generative methods for longitudinal modeling, patient simulation and medical digital twins. |
| August 2026 | Completed Google Summer of Code 2026 with the Python Software Foundation, contributing new registration methods, performance improvements, bug fixes, tests and documentation to DIPY. Read the final project report. |
| June 2026 | Our paper DesigNet received the Best Full Paper Award at CEIG 2026. The work presents a structured approach to generating editable vector graphics. |
| May 2026 | DesigNet was accepted for publication in Computers & Graphics and for presentation at CEIG 2026. |
| October 2025 | Gave the talk “Efficient Transformers: Training & Inference on a Budget” at Madrid Innovation Lab, organized by Machine Learning Circle. |
| July 2025 | Defended my master’s thesis, Generative Font Shape Modeling, at Universidad Politécnica de Madrid and received a grade of 10/10. |
| June 2025 | Received the José Cuena Award, an annual distinction for incoming full-time students with the highest average across the seven first-semester courses of UPM’s MSc in Artificial Intelligence. |