Ferran Espuña

Madrid, Spain
ferranespuna@gmail.com · +34 600 24 69 87 · GitHub

Research Interests

Arithmetic combinatorics; higher-order Fourier analysis; additive combinatorics; harmonic analysis; ergodic theory; extremal combinatorics.

Current Position

PhD Student in Mathematics

Universidad Autónoma de Madrid (UAM) / Instituto de Ciencias Matemáticas (ICMAT)
2026–Present

Supervisor: Pablo Candela

Research area: arithmetic combinatorics, with emphasis on higher-order Fourier analysis, Gowers norms, inverse theorems, nilspaces, and additive structures.

Education

M.Sc. in Advanced Mathematics and Mathematical Engineering

Universitat Politècnica de Catalunya (UPC)
2025
GPA: 9.45/10

Master’s thesis: Finding Partite Hypergraphs Efficiently.

Relevant coursework: Commutative Algebra, Number Theory, Coding Theory, Cryptography, Combinatorics, Graph Theory, Computational Complexity.

B.Sc. + B.Sc. Double Degree in Mathematics and Computer Science

Universitat de Barcelona (UB)
2023
GPA: 9.0/10
Extraordinary Bachelor’s Degree Award.

Publications

Espuña, Ferran. “Finding Partite Hypergraphs Efficiently.” Information Processing Letters, 2026.
DOI · arXiv

Palomar, Jorge, et al. “A CURATEd CATalog: Rethinking the Extraction of Pretraining Corpora for Mid-Resourced Languages.” Proceedings of COLING-LREC 2024.
Paper

Research Experience

Research Engineer

Barcelona Supercomputing Center (BSC), Language Technologies Unit
2023–2026

Research on large language models, multilingual pretraining corpora, model evaluation, and scalable AI systems.

Selected contributions:

  • Development and automation of CURATE, a large-scale corpus processing pipeline for HPC environments
  • Contribution to CATalog, a large Catalan pretraining corpus
  • Contribution to the Salamandra multilingual language model family
  • Research on open-ended generation evaluation, state space models, sparse autoencoders, and multimodal architectures

Research Intern

Computer Vision Center (CVC)
2022

Project on topological data analysis methods for the study of neural network generalization.

Supervisors: Sergio Escalera, Carles Casacuberta, Rubén Ballester.

Research Assistant, Image Processing

ChipScope Research Group, Universitat de Barcelona
2022

European project on chip-scale microscopy. Work included computational imaging pipelines, alignment algorithms, and wave backpropagation.

Technical Skills

Mathematical Areas: arithmetic combinatorics, graph theory, discrete mathematics.

Machine Learning / HPC: PyTorch, Hugging Face Transformers, distributed training, Slurm, Docker, Linux.

Programming: Python, C/C++, Java, Bash.

Languages: Spanish (native), Catalan (native), English (C2).

Awards and Distinctions

  • Extraordinary Bachelor’s Degree Award, Universitat de Barcelona
  • Stanford University, Machine Learning Specialization
  • IELTS Academic: 8.5/9

Selected Project

Complex fractal shaders.
Interactive GLSL fragment shader visualising fractals emerging from complex dynamical systems.