Hi, I'm David 👋
MSc student in Health Data Science (URV, in collaboration with UB, UAB, UPC, UdG, UdL, UVic-UCC & Grenoble Alpes) | BSc Biomedical Engineer & BSc Industrial Electronics and Automatic Control Engineer (UPC, Barcelona)
I work at the intersection of biomedical engineering and applied data science: signal processing, clinical data pipelines, and ML/AI systems for healthcare — with hands-on exposure to cloud infrastructure (AWS) and reproducible computing (Snakemake, SLURM, Docker).
-
🔬 Currently: clinical data analytics on ICU signal data, co-author on an accepted IEEE EMBC 2026 paper
-
🧠 Interests: clinical NLP & EHR, medical imaging, survival analysis, MLOps for health data, cloud architecture for regulated environments
-
📍 Based in Barcelona, Spain
-
📫 Reach me: hidalgo.fabregas.david@gmail.com · LinkedIn
Python R SQL scikit-learn PyTorch TensorFlow NetworkX · AWS (EC2, S3, Lambda, VPC) Docker SLURM Snakemake · MIMIC-III DICOM ECG/PPG
-
CHD Risk Classifier FastAPI-Docker (team project — forked, my contribution highlighted in the README) — End-to-end ML pipeline classifying 10-year coronary heart disease risk on a 4,000+ patient cohort; I led the FastAPI service design and Docker deployment, and coordinated the team's technical review.
-
HPC SLURM Job Orchestrator — Python orchestrator for dependent, distributed job executions on a SLURM-managed HPC cluster: job arrays, dependency chaining, and automated result consolidation.
-
BioInformatics Sequence Analysis — Reproducible bioinformatics pipeline (FASTA parsing, DP sequence alignment, CIGAR analysis) orchestrated end-to-end with Conda + Snakemake.
-
CS50AI — Portfolio of Harvard CS50 AI problem sets (search, knowledge/logic, Bayesian inference, CSPs, RL, CNNs, NLP), each documented with design decisions and results.
-
ATP Match Predictor (🚧 work in progress) — Random Forest model estimating a player's win probability, compared against market-implied probabilities from bookmaker odds to flag pricing discrepancies; leak-free walk-forward validation and dynamic feature engineering. (rolling form, surface-aware H2H).
(See pinned repos below for details, code, and READMEs.)