Mahmoud Youssef
AI researcher working on privacy-preserving synthetic data generation at Wrocław University of Science and Technology.
I am a researcher in privacy-preserving generative AI. My work sits at the point where three things meet: differential privacy, which decides what a model is allowed to learn; generative modelling, which decides how well it learns it; and trustworthy machine learning, which asks whether the result can be relied on once someone acts on it. My doctoral research at Wrocław University of Science and Technology extends this to data that arrives as a stream and refuses to hold still.
I did not start in research. Before Wrocław I spent well over a year as a data analyst at Advanced Pipeline Services in Abu Dhabi, building automated pipelines over large multi-modal sensor datasets and designing SQL systems with role-based access control for industrial field operations. It was unglamorous work and it shaped how I think. Handling data that is genuinely sensitive, where access has to be justified, where a bad record has physical consequences downstream, is a different experience from reading about privacy as an abstraction. The question I ended up working on, how to use data responsibly when you cannot simply look at it, came from there rather than from a paper.
My master's thesis, Privacy-Preserving Synthetic Data Generation for Pattern Classification Tasks, developed methodologies for generating tabular synthetic data that keeps formal privacy guarantees without destroying downstream classifier performance, benchmarked against PATE-GAN, DP-CTGAN and DP-GAN. It produced PATE-TabTransGAN, accepted at the SYNDAiTE workshop at ECML PKDD 2026. The PhD continues that line of work in a harder setting: multi-modal data streams, non-stationary distributions, and continual learning. What I care about most in all of it is decision faithfulness, whether a model trained on synthetic data actually decides the way the real data would have.
Education
-
Oct 2026 – Present
PhD, Information and Communication Technology
Wrocław University of Science and Technology · Wrocław, Poland
Research topic: Privacy-Preserving Synthetic Data Generation for Trustworthy and Adaptive Machine Learning. Designing decision-faithful and privacy-preserving synthetic data generators across multi-modal data streams, under non-stationary distributions and continual learning setups.
-
Mar 2025 – Jul 2026
MSc, Applied Computer Science
Wrocław University of Science and Technology · Wrocław, Poland
Thesis: Privacy-Preserving Synthetic Data Generation for Pattern Classification Tasks. Developed novel methodologies for generating tabular synthetic datasets that preserve privacy guarantees while maintaining downstream classifier performance, with a comprehensive empirical comparison against PATE-GAN, DP-CTGAN and DP-GAN. GPA 4.6 / 5.0.
-
Oct 2018 – Sep 2022
BSc, Computer Science
Damietta University · Damietta, Egypt
Foundations in algorithms, data structures, linear algebra, statistics and machine learning. Capstone project: a penetration-testing toolkit for network reconnaissance, automating asset discovery, port scanning and service fingerprinting.
Professional background
-
Mar 2023 – Jul 2024
Data Analyst
Advanced Pipeline Services Limited, LLC · Abu Dhabi, UAE
Built and integrated automated data analysis pipelines in Python over large-scale multi-modal operational and sensor datasets, supporting feature extraction and time-series analysis for infrastructure monitoring. Designed secure SQL database systems with role-based access control, data expiry alerting and integrity constraints for sensitive industrial data streams. Also engineered a panorama playback application compatible with heterogeneous camera systems, automating visual data review in field operations.
The relevance to what I do now is direct: multi-modal sensor streams, controlled access to sensitive data, and the practical reality that a data system's failures are somebody's problem.
Technical background
Awards
-
Outstanding Achievement
Wrocław University of Science and Technology
Awarded at the graduation ceremony in recognition of outstanding academic performance and an exceptional thesis and research contribution.
-
Rector's Scholarship
Wrocław University of Science and Technology
Merit-based award for outstanding academic performance during the MSc in Applied Computer Science.
Get in touch
I am glad to hear from researchers working on privacy, synthetic data, or trustworthy machine learning, whether that is a question about a paper, a possible collaboration, or an idea worth arguing about. Based in Wrocław, Poland.
MahmoudGamalfl8@outlook.com