Dr. Ozgur Ural
Dr. Ozgur Ural

U.S. Ph.D. (Embry-Riddle) | ML research scientist & senior software engineer | Architect of mission-critical systems

Verifiable & Trustworthy ML · Applied GenAI · Mission-Critical Systems

Verifiable machine learning.
Reliable systems under failure.

I'm Dr. Ozgur Ural, a machine-learning research scientist and senior software engineer at Avion. I research proof-of-learning, model watermarking, and adversarial robustness, and build real-time flight-simulation platforms.

Leiden, Netherlands. Open to technical advisory, speaking, and research collaboration.

Read my research Advisory & speaking Get in touch
  • 12+Years in mission-critical software
  • Ph.D.ML research · ERAU (USA) 2025
  • 5First-author papers · 3 in IEEE Access

Leadership & Core Domains

I work where hands-on engineering meets organizational scale: setting architectural direction, raising engineering standards, and keeping complex systems, from mission-critical C++ to cloud-native microservices, correct under load and under failure.

  • Engineering Leadership

    Direction, standards, teams

    Having shipped national data-leakage-prevention products and mission-critical software, I now set architecture decisions and engineering standards across the organization: resilience patterns, review culture, and release discipline.

  • Fault-Tolerant Distributed Systems

    Correct when parts fail

    Real-time simulation platforms and cloud-native services (C++, Scala, TypeScript, gRPC) built on distributed protocols whose outputs and state can be audited end-to-end without compromising latency.

  • Trustworthy ML & Security

    Securing AI pipelines

    My Ph.D. and three first-author IEEE Access papers defend ML training integrity against spoofing attacks, through feature-based model watermarking and proof-of-learning verification.

  • Adversarial Robustness

    Models that survive adversaries

    I test where models break: provenance verification, adversarial examples, and the limits of a claimed model identity. A model earns trust by surviving contact with an adversary, not by scoring well on a clean test set.

Timeline

  1. 2026 Authored Avion's AI adoption roadmap and an enterprise reference architecture covering on-premises inference, governed cloud APIs, and edge workloads. Built the platform presenting the strategy and developed engineering-knowledge retrieval prototypes. Projects distinguishes implemented work from proposed agent workflows.
  2. Dec 2025 Published SecurePoL: Integration of Watermarking with Proof-of-Learning to Enhance Security Against Spoofing Attacks in IEEE Access. Read paper.
  3. Dec 2025 Joined the Program Committee for the 2nd Workshop on NLP Applied to Information and Cyber Security (NLPAICS 2026), University of Alicante. Conference.
  4. Aug 2025 Conferred Ph.D. in Electrical Engineering & Computer Science (ERAU). Dissertation: Enhancing Proof-of-Learning Security Against Spoofing Attacks Using Model Watermarking. My doctoral research was guided by Dr. Kenji Yoshigoe (Committee Chair) and IEEE Fellow Dr. Houbing Song, with whom I continue to actively collaborate on securing distributed ML systems. Dissertation · Verify diploma.
  5. 2021–2025 Graduate Research Assistant at Embry-Riddle Aeronautical University. Three IEEE Access publications on proof-of-learning and model watermarking.
  6. Nov 2024 First-author paper Feature-Based Model Watermarking for PoL in IEEE Access. Read paper.
  7. Dec 2023 Published the survey Blockchain-Enhanced Machine Learning in IEEE Access. Read paper.
  8. Oct 2023 Joined Avion Full Flight Simulators as Senior Software Engineer. Architecting real-time simulation platforms and cloud-native infrastructure for Level D flight simulators.
  9. May 2021 Published Automatic Detection of Cyber Security Events from Turkish Twitter Stream and Newspaper Data at ICISSP, co-authored with my M.Sc. advisor Prof. Cengiz Acartürk. Read paper.
  10. 2020–2021 Software Team Lead at Havelsan. Led a 14-engineer team delivering the Havelsan DLP data leakage prevention product for defence and government clients.
  11. 2019–2020 Expert Software Engineer at STM Defence Technologies. Developed critical mission-control and ground-control systems for Kargu, Turkey's first indigenous loitering-munition (kamikaze) UAV and an internationally recognized autonomous system, alongside the Togan reconnaissance UAV. Hard-real-time and safety-critical, where a missed deadline is a mission failure.
  12. 2014–2019 Expert Software Engineer at Comodo Cybersecurity. Led design and architecture of the Secure Web Gateway, enterprise Patch Manager, and the Chromium-based Dragon browser.
  13. 2014 Graduated METU Computer Engineering with two competition wins: 1st prize at the national project competition of TÜBİTAK (Turkey's national science and technology council), and 2nd prize at the METU Computer Engineering (CENG) graduation Demo Day.
  14. 2010–2014 B.Sc. in Computer Engineering at Middle East Technical University (METU), Turkey's top-ranked and most selective computer engineering program (English-medium): admission among roughly the top 1,000 of more than 2 million national university-entrance candidates.

Selected Work

Scientific Impact & Professional Service

My research examines what evidence can support claims about a model’s training history and ownership, and how those claims can fail under attack. The publication record links to the papers, methods, and reported evaluations.

I review manuscripts and conference submissions across machine learning, security, and distributed systems:

  • Journal Referee, IEEE Access, and other peer-reviewed journals in ML security, privacy, and distributed systems
  • Program Committee Member, 2nd Workshop on NLP Applied to Information and Cyber Security (NLPAICS 2026), University of Alicante
  • Reviewer, IEEE and ACM conferences in security, machine learning, and distributed systems

Get in touch

For research collaboration, technical advisory, or speaking engagements, reach out by email, or browse my publications, projects, field notes, and essays.