🎯 AI Engineer | 🤖 Software Developer | 🎲 Boardgame & RPG Enthusiast

Passionate about AI, software engineering, puzzles and game design, I thrive at the intersection of innovation and real-world impact, tackling new challenges to create solutions that inspire and bring wonder to the world.

Research on integrating Style Transfer into an Interactive Neural Painting Transformer, enabling AI-driven brushstroke suggestions that guide users in painting a scene while matching a reference artistic style.

📄 Master's Thesis -- UniTN MHUG research group

Research on implementing Overlapping Community Detection algorithms in graphs, enhancing Community Identification through improved label propagation techniques and performance evaluation metrics.

📄 Bachelor's Thesis

A serverless Python-powered Telegram bot @iactabot for rolling dice, supporting advanced expressions and damage types, currently running on AWS Lambda.

A Flutter mobile app for managing large-scale Scout events (used during the Challenge 2019 provincial event), enabling real-time tracking of patrols and checkpoints positions. Features a real-time map, a scoreboard management system, secure authentication, and location sharing. Powered by REST APIs on a PHP backend on a MySQL database.

The npm TypeScript library enumify-ts is a fork of enumify that fully supports generics, enabling correct type inference for collections.

A machine learning project for classifying thoracic X-ray images and detecting illnesses such as Viral Infection, Bacterial Infection, and COVID-19. The project compares both traditional machine learning models and deep neural networks for accurate predictions.

📄 Machine Learning

A recommendation system designed to suggest personalized therapies for patients based on the similarities with other patients and conditions. The system utilizes a three-way relationship between patient, condition, and therapy. It is built on a synthetic dataset and includes collaborative filtering and similarity-based methods for improved therapy recommendations.

📄 Data Mining

A single deep learning model for multi-class, multi-label classification of visual attributes and person re-identification. It classifies characteristics such as age, accessories, clothing, and gender, and uses an embedding-based re-identification

📄 Advanced Computer Vision

Evaluation of various Graph Neural Network (GNN) architectures and proposal of data augmentation properties to enhance their performance across tasks like node classification, link prediction, and graph classification.

📄 Advanced Topics in Machine Learning and Optimization

A computer vision system that tracks players and ball possession changes in basketball videos using a YOLO detector and a CSRT tracker.

📄 Computer Vision

A Python educational toolkit for extracting and analyzing fabric texture features using statistical methods like Autocorrelation Function, Local Binary Patterns, and Co-Occurrence Matrices.

📄 Signal, Image and Video

📧 Email: matteo.zanella98@gmail.com
💼 LinkedIn: linkedin.com/in/matteozanella
🎮 Gaming & Boardgames: Always up for a match!