Dr. Ozgur Ural
I do machine learning research and write software that cannot fail.
I'm Dr. Ozgur Ural. My doctoral work is in machine learning: published research on proof-of-learning, model watermarking, adversarial robustness, and blockchain-enhanced ML. Four IEEE Access papers (2023 to 2025); Ph.D. dissertation, ERAU 2025. By day I'm a senior software engineer building production systems where bugs aren't allowed to exist. The parts of AI I keep returning to are the ones where the math has to survive contact with adversaries and reality.
Currently in Leiden, Netherlands. Writing, building, publishing.
- 11+Years engineering
- Ph.D.ERAU · 2025
- 4IEEE Access papers
- PCNLPAICS 2026
Research focus
Published research is concentrated in trustworthy ML: the question of when a model can actually be trusted, and how to prove it. Broader interests stretch across machine learning: evaluation, robustness, alignment, agent verification, and the engineering practices that make ML research reproducible and reliable.
- Trustworthy ML
Proof-of-Learning & watermarking
Defending ML training integrity against spoofing attacks. Feature-based model watermarking integrated with proof-of-learning. IEEE Access 2024, 2025. Ph.D. dissertation, ERAU 2025.
- Adversarial & Robust ML
When models can be trusted
Adversarial examples, provenance verification, defences against spoofing. The broader question of when a model's claimed identity and behaviour can be trusted at all.
- Distributed & Verifiable Systems
Auditable ML pipelines
Blockchain-enhanced ML (IEEE Access 2023 survey), cloud-native services in Scala / TypeScript / gRPC, and the design of distributed protocols whose outputs can be audited end-to-end.
- Engineering for ML
Research code that works in production
Real-time, mission-critical C++ in environments where bugs are not tolerated. The engineering discipline that ML research code mostly lacks, and that AI labs increasingly need.
Recent
- Dec 2025 Joined the Program Committee for the 2nd Workshop on NLP Applied to Information and Cyber Security (NLPAICS 2026), University of Alicante. Conference.
- Dec 2025 Published SecurePoL: Integration of Watermarking with Proof-of-Learning to Enhance Security Against Spoofing Attacks in IEEE Access. Read paper.
- Aug 2025 Conferred Ph.D. in Electrical Engineering & Computer Science (ERAU). Dissertation: Enhancing Proof-of-Learning Security Against Spoofing Attacks Using Model Watermarking. Dissertation.
- Nov 2024 First-author paper Feature-Based Model Watermarking for PoL in IEEE Access. Read paper.
- Dec 2023 Published the survey Blockchain-Enhanced Machine Learning in IEEE Access. Read paper.