About Me

Welcome! I’m Diego Villalba, a BSc student in Physics at the Universidad Nacional Autónoma de México, currently specializing in cosmology, computational astrophysics, and machine learning.

I am deeply interested in the intersection of machine learning and cosmology, particularly in how modern ML methods can enhance the analysis and reconstruction of cosmological simulations. Currently, during my research stay at the Donostia International Physics Center (DIPC), I am expanding this line of work by studying how differentiable models, neural implicit representations, and simulation-based inference can be used to analyze, compress, and reconstruct cosmological fields with higher precision and physical fidelity. At UNAM, I worked on ML-based super-resolution of cosmological fields, exploring techniques to upscale and recover high-resolution structure from coarse N-body data.

In addition to cosmology, I work on machine learning–based solutions, both in:

• Medical imaging — as a side project focused on MRI acceleration, leveraging neural reconstruction techniques to reduce acquisition time while preserving diagnostic quality.
• Industry — working on implementations of OCR methods applied to industry pipelines

Recently, I had the opportunity to work as a research student at the Donostia International Physics Center (DIPC) in San Sebastián, Spain, under the supervision of Dr. Raúl Angulo and Dr. Carolina Cuesta-Lázaro. There, I worked on projects involving compressed representations of cosmological simulations, displacement-field reconstruction, and ML-driven approaches to high-resolution structure formation. This experience strengthened my interest in computational cosmology and gave me hands-on exposure to cutting-edge research using high-performance computing, PyTorch-based ML models, and next-generation simulation tools.

Throughout my academic journey, I’ve taken part in multiple research projects, attended specialized workshops, and collaborated in student-led scientific initiatives. These experiences have deepened my curiosity about how machine learning can transform computational modeling across domains — from cosmology and astrophysical data analysis to biomedical imaging.

Research Interests

• Machine Learning for Cosmology (super-resolution, simulation-based inference, differentiable models)
• ML-based Super-Resolution for MRI Acceleration
• N-body and hydrodynamical simulations
• High-performance and parallel computing
• Cosmological structure formation and large-scale dynamics

Personal

I’m originally from Mexico City. Outside of research, I enjoy tutoring new physics students, working in teaching labs, astrophotography, and participating in student organizations and science-communication activities. I strongly believe in making science accessible and building a collaborative academic environment.

I’m always excited to connect with people who share interests in cosmology, ML, scientific computing, or data-driven science, so feel free to reach out!

Explore my projects, coursework, blog posts, and portfolio, and don’t hesitate to get in touch if you’d like to chat or collaborate.