I'm a Master's student in Data Science at EPFL, on an Excellence Fellowship (ESOP, top 0.004% of incoming MSc students). I work at the EPFL Computer Vision Lab on transformer architectures for 3D scene reconstruction, and I'm currently building a real-time controllable text-to-motion model with kinematic conditioning.
I graduated in Artificial Intelligence from the Autonomous University of Barcelona with the highest GPA across all engineering degrees (9.41/10). I have 5 papers at CVPR / ECCV (4 as first author), specializing in generative AI and controllability in vision models. This past summer I was a Machine Learning intern at GoodNotes in London, where I shipped a diffusion-based text-to-SVG feature now used by millions.
Controllable Autoregressive Text-to-Motion Generation 🔗
EPFL CS-503 (in progress)
Real-time text-to-motion model that lets the user steer generation mid-sequence: at any frame, specify a full-body pose, hand or foot position, or a path the character must follow. Built on T2M-GPT + VQ-VAE, trained on BONES.
LayeredDoc: Domain Adaptive Document Restoration 🔗
ECCV WiCV 2024, ICDAR 2024
Towards Controllable Image Relighting 🔗
Bachelor's dissertation 2025
Student Researcher at EPFL Computer Vision Lab
Lausanne, Switzerland
October 2025 - Present
Machine Learning Intern at GoodNotes
London, United Kingdom
July 2025 - September 2025
Undergrad Researcher at Computer Vision Center
Barcelona, Spain
September 2023 - June 2025
Master's in Data Science
EPFL, Laussane, Switzerland
September 2025 - August 2027
Bachelor's in Artificial Intelligence
Autonomous University of Barcelona, Barcelona, Spain
September 2021 - June 2025
MSc Data Science at EPFL on a top 0.004% Excellence Fellowship. Graduated UAB with the highest GPA across all engineering degrees (9.41/10), with honors on 18 courses including 3D Vision, Generative Models, Deep Learning, CV and NLP.