Work

I have worked on the technical side of privacy since 2005, alternating between research and the market. This duality is key: my research background shows me where the real questions are decided, and my years in the market taught me that a solution must work on time and within budget.

How I can help

• Anonymisation, identification and de-anonymisation

  What makes a piece of data or a person identifiable, and how strong is the anonymisation applied?

  • MNB: anonymity of transaction data

• Web tracking and device fingerprinting

  Recognising and linking users and devices, and its privacy and fraud prevention aspects.

  • SEON: device scoring

• Face recognition and privacy-preserving video analytics

  Biometric identification, reversibility of face recognition fingerprints, and pseudonymised solutions.

  • VeriDome

• Technical privacy expert opinion and privacy by design

  Technical assessment, risk evaluation and privacy-focused design of systems and system plans.

  • Privacy impact assessment from a technical angle (slides)

• Privacy of AI systems

  Privacy risks of machine learning, generative AI and automated processes, and their secure implementation.

  • Machine learning methods for de-anonymisation

If any of these sound familiar, write to me.

In addition, I regularly give talks and internal trainings on these topics and beyond. I also give interviews to the press — see more on the Press page.

Where I have worked

The periods often overlap.

Vitarex Stúdió Kft. — managing director (since 2023), previously project manager (from 2018). In addition to management tasks, I am especially proud of two projects. In the VeriDome) project, since 2023 we have developed a face recognition procedure with an additional privacy layer, and based on this, created a complex CCTV analytics solution applicable in several industries. The IML4E) project (2021–2024) focused on the industrial application of machine learning solutions; in this, Vitarex, as leader of the Hungarian consortium, developed the Tartás Pajtás) tool to assist scoliosis screening. In 2025, the project won the ITEA Award of Excellence for Standardisation) award.

Sole proprietor — web development (2009–2023). I mainly built websites; this was an excellent way to gain a deeper understanding of web technologies. The practical knowledge I gained here became decisive in my later research and development work, especially in the field of web tracking and identification.

SEON — device fingerprint team lead and researcher (2017–2022). I had the opportunity to build a great R&D team, with whom we created world-class fraud detection solutions, for example device scoring with dynamic components.). Many of these solutions are still used by companies in Hungary and abroad.

BME-AUT — research fellow (2019–2021). My main topics were privacy in face recognition and the study of generative machine learning methods.

Inria (France) — postdoctoral researcher and research engineer (2015–2018). My main projects examined the uniqueness and identifiability of web users; we presented this in the European Parliament.

BME-HIT — PhD student, then research assistant (2007–2015). De-anonymisation of networked datasets and defences against it — this was also the topic of my doctoral dissertation — as well as privacy issues of web tracking.

Teaching

Between 2007 and 2015, then again between 2019 and 2021, I taught regularly at the Faculty of Electrical Engineering and Informatics at BME. Thirty-three theses were completed under my supervision: eleven of their authors won thesis awards, two won TDK awards, and seven scientific publications resulted from their work. In spring 2012, I launched my own course "Privacy-enhancing technologies" (privacy technologies, VIHIAV00), and between 2019 and 2021 I organised a machine learning talent programme at BME-AUT. I taught the Management of personal and public data) course, and regularly gave guest lectures in the Privacy and freedom of information) course.

These are just the main milestones. At the AUT department I was also responsible for other courses, and in the HIT and AUT programmes I regularly gave guest lectures in at least a dozen subjects.

Supervised students and their results