Applied Computer Science graduate with a specialization in Machine Learning and Big Data.
My work spans federated weather forecasting, GPU computing, and embedded systems research. I've been tinkering with computers since childhood, and I still enjoy figuring out how things work, experimenting with code, and exploring unfamiliar areas of computing.
FederatedWeatherCC · Federated weather forecasting
Collaborative research on forecasting from distributed weather-station data using Transformer and Crossformer models. The repository includes data preprocessing, centralized training, and federated client/server code built with Python, PyTorch, and Flower. I contributed substantially to the research behind the accompanying BigHPC paper, which credits three authors equally and names me as corresponding author.
Samsung Tizen 4 USB root · Embedded systems & reverse engineering
A project I'm especially proud of: I adapted Luminger's USB-root research to Samsung's 2018 Kant-M2 platform and validated an interactive root shell on a QE49Q6FNA running firmware T-KTM2DEUC 1400.2, without modifying the firmware. The repository documents the platform-specific analysis and tooling.
JCAM_Cuda · GPU computing for biological simulation · Paper
I ported an existing MATLAB implementation to CUDA for a numerical model of tumor angiogenesis. The published work implements GPU kernels for pivot selection, row swapping, and Gaussian elimination. It reports up to 113× speedup on an NVIDIA H100 against the sequential CPU implementation in the paper's benchmarks.
Esercitazioni-C-Parthenope · Learning resources in C
A collection of programming exercises for fellow students at the University of Naples Parthenope, covering sorting, text processing, set operations, and exam practice.
- Federated Learning for Distributed Weather Forecasting: A Practical Approach on Real Multidimensional Georeferenced Data — BigHPC 2025, co-located with ITADATA 2025; CEUR Workshop Proceedings, Vol. 4124. Equal-contribution author and corresponding author.
- A GPU-CUDA Numerical Algorithm for Solving a Biological Model — AppliedMath, 5(4), 178, 2025. My contribution: software implementation, including the MATLAB-to-CUDA port.
- Machine learning: Python, PyTorch, Flower; time-series forecasting and federated learning.
- Scientific computing: CUDA, MATLAB, C; numerical algorithms and GPU parallelization.