Hi! I'm
MSc student at the University of Trento, specializing in Artificial Intelligent Systems. Currently researching VQA & Scene Graphs at Polytechnic University of Milan's AIRLab.
About Me
Hey! I'm Donald — a software engineer and AI researcher with a passion for building intelligent systems that bridge the gap between research and real-world applications.
I'm currently pursuing my MSc at the University of Trento, where I focus on machine learning, computer vision, and NLP. My thesis work at Polytechnic University of Milan's AIRLab explores Visual Question Answering and Spatio-Temporal Scene Graphs.
Before grad school, I earned my BSc in Computer Science from Ca' Foscari University of Venice, where I worked as a Research Assistant on ML model watermarking — work that resulted in a publication at EDBT 2025.
Outside of academia, I work part-time at Cluster Reply as a backend developer, and I enjoy competitive programming on Codeforces where I've achieved the rank of Specialist.
I'm always open to research collaborations, interesting engineering challenges, and conversations about AI. Feel free to reach out!
Programming Languages
Areas of Interest
Experience
Part-Time Backend Developer
Aug 2024 — PresentCluster Reply
- Developed C# customizations to deliver client-requested features, reducing manual effort by 15+ hours per week through automation of key processes.
- Resolved functional and performance bugs in test and production 30% faster than role-level expectations using debugging, logs and SQL.
- Collaborated on 4+ D365/AX projects, focusing on Finance, Supply Chain, and Manufacturing modules, leveraging C# and SQL.
Research Assistant
Feb 2024 — Jul 2024Ca' Foscari University of Venice
- Co-authored verifiable ownership of tree-ensemble models by implementing a trigger-set watermark with only a 2% accuracy overhead.
- Enhanced watermark robustness against removal attacks, leading up to 25% reduction in attack success rates.
- Established anti-forgery bounds using Z3 SMT solver; the best counterfeit achieved 14% of the genuine signal and reduced test accuracy by 30–40 percentage points.
Education
Thesis Research — VQA & Spatio-Temporal Scene Graphs
Feb 2026 — PresentPolytechnic University of Milan — AIRLab
Working on Visual Question Answering and Spatio-Temporal Scene Graphs.
MSc in Computer Engineering — Artificial Intelligent Systems
Sep 2024 — PresentUniversity of Trento
GPA: 3.8 / 4.00
Relevant Coursework
BSc in Computer Science
Sep 2021 — Jul 2024Ca' Foscari University of Venice
GPA: 3.9 / 4.00
Relevant Coursework
Projects
Selected work from research, university, and side projects.
Publications
Watermarking Decision Tree Ensembles
EDBT 2025Stefano Calzavara, Lorenzo Cazzaro, Donald Gera, Salvatore Orlando
28th International Conference on Extending Database Technology (EDBT 2025) — Barcelona, Spain — Mar 25–28, 2025
A novel approach to verifiable ownership of tree-ensemble models through trigger-set watermarking.
Get in Touch
I'm always open to interesting conversations, research collaborations, and new opportunities. Feel free to reach out through any of the channels below.