About

Hi! I'm Pau Boncompte Carré, a Biomedical Engineer from Barcelona currently pursuing a Master’s in AI at Institute of Science Tokyo. I work with medical imaging, neural data, and clinical datasets, building AI-driven tools that support data-driven decision-making and improve diagnostic and research workflows.

In my free time, I'm usually reading, playing videogames, or learning languages (or at least trying to!).

Experience

AI Intern

Siemens Healthineers Japan | Aug. 2026 – Present

  • Developing an AI agent for application architecture design.

Data Science Intern

MI-6株式会社 | Jul. 2026 – Present

  • Developing an CV pipeline for dental condition classification from intraoral images. The conditions detected span from caries to fluorosis and MIH.

Graduate Research Student

Yoshimura Lab, Institute of Science Tokyo | 2024–Present

  • Designed a VR environment in Unity 6 with real-time LSL-synchronized 64-channel EEG acquisition for locomotion and turning motor imagery.
  • Developed a hierarchical deep learning ensemble (EEGNet + ShallowConvNet in Python/TensorFlow), achieving 87% binary and 63% four-class accuracy for locomotion decoding.

Biomedical Engineering Intern

SIMBIOSys, BCN Medtech | Nov. 2023 – Aug. 2024

  • Analyzed MRI neuroimaging data from Hospital Clínic de Barcelona comparing cortical and cerebrovascular structure between healthy controls and patients born with intrauterine growth restriction (IUGR)
  • Performed comparative analysis to identify structural differences in brain morphology and vascular architecture associated with IUGR

Education

Master’s in Artificial Intelligence – Institute of Science Tokyo | 2024–Present
Bachelor’s in Biomedical Engineering – Universitat Pompeu Fabra | 2020–2024

Skills

  • Programming: Python, MATLAB, C#, SQL, Java, Bash, Lua
  • Frameworks & Tools: Pandas, PyTorch, TensorFlow, Scikit-learn, NumPy, Matplotlib, pydicom, MATLAB, FreeSurfer, Docker, Git, PostgreSQL, Linux, AWS, HPC environments
  • ML Engineering & MLOps: ML pipelines, hyperparameter optimization, data augmentation, HPC/SLURM job scheduling, AWS, HPC environments
  • Languages: Catalan and Spanish (Native), English (Fluent), Japanese (N2, Business-level), French (B1) and German (Beginner).

Publications

  • Boncompte-Carre, P., Martinez-Tejada, L. A., & Yoshimura, N. (2026). EEG-based forward movement and turning MI classification with and without action observation in virtual reality. 2026 14th International Conference on Brain-Computer Interface (BCI), 1–6. DOI: 10.1109/BCI69045.2026.11435097.

About this Website

This website has been created with Pelican, a static site generator powered by Python. The theme used is Flex by Alexandre Vicenzi, with a few additional scripts I added to enable grid-based image visualization, lightboxes, and mobile-adapted UI. You can check out the source code on my GitHub.