I'm a mathematical engineer specialised in data science and a telecommunications engineer, currently pursuing an MSc in Finance at ESADE as a recipient of the European Talent Scholarship.
This profile is essentially my engineering notebook: most of what you'll find here comes from my double degree at Universitat Pompeu Fabra, plus a few hackathon and side projects. I've published it because the work explains my background better than a CV bullet does — you can see how I actually approach a problem, not just which courses I passed.
Neural networks built from scratch and applied end-to-end.
- deep-learning-pytorch-fundamentals — a NumPy MLP with hand-derived backprop and a momentum optimiser, then re-implemented in PyTorch
- deep-learning-cnn-transfer-learning-svhn — CNN architecture comparison, a 132K-parameter budget model, and transfer learning on street-view digits
- bank-customer-intelligence-ml — customer segmentation, loan-approval classification with ensembles, and a from-scratch MLP
- Massive_Data_Mining — near-duplicate detection, time series forecasting, recommender systems
- Stock_Price_Prediction — deep learning for price prediction and trend capture
Models are only useful if someone can act on them.
- pancreatitis-risk-shap-app — clinical risk prediction with SHAP explanations, because a black box isn't much use at a bedside
- streamlit-salary-explorer — exploring and predicting developer salaries from survey data
- Ecommerce_Product_Pricing_Recommendation — trend identification and price optimisation for online sellers
- innosuisse-startupticker-gingershots — SwissHacks 2025: AI-driven insights for Swiss startups and investors
The theory, implemented rather than just described.
- bayesian-networks-belief-propagation — sum-product belief propagation written from scratch and validated against pgmpy
- statistical-models-matlab — MANOVA via Wilks' Lambda, bootstrap resampling, PCA, ISOMAP, Kernel PCA
- network-science-problem-sets — PageRank, HITS, random-graph models, spectral analysis
Where the telecom degree shows up.
- wireless-comm-cellular-simulator — Monte Carlo simulation of a 19-cell network: frequency reuse, fractional power control, SIR and throughput
- wifi-antenna-design-nec — four 2.4 GHz antennas modelled in NEC2, optimised, then 3D-printed and measured
- optical-access-metro-network-design — FTTH/GPON fibre design for three rural Catalan municipalities
- priority-queueing-simulator-cost — discrete-event simulation of priority queueing under bursty arrivals
- dhcp-attribution-layer-incident-response — resolving "which host had this IP at time T" for incident response
- operating-systems-c-labs — POSIX C: fork/exec, pthreads, producer/consumer
- parallel-programming-openmp-mpi-openacc — OpenMP, MPI and OpenACC with real cluster output
- Large-Scale-Distributed-Systems — Java, AWS and Spark
- Search_Engine_IRWA / Information_Retrieval_Web_Analysis — a search engine, backend through web app
- Music_Festival_SQL_Database — SQL design for large-scale festival data
- system-dynamics-service-desk-vensim — stock-and-flow modelling of a three-tier IT service desk
This list keeps growing — I add projects as I finish them, and new categories will appear as my work moves further into finance and quantitative modelling.
Languages Python · C · C++ · Java · MATLAB · SQL ML PyTorch · TensorFlow/Keras · scikit-learn · SHAP Data & apps pandas · NumPy · Streamlit · Plotly · NetworkX Systems OpenMP · MPI · OpenACC · Spark · AWS · PostgreSQL Domain NEC2 · QGIS · Vensim
Much of this is team coursework, and I've tried to credit collaborators in the individual READMEs — the good ideas in these repos were rarely mine alone.
Always happy to connect and exchange ideas 🤝 LinkedIn