I am an Assistant Professor in Computer Science (Data Management for AI) at the University of Birmingham, The UK. Previously, I was a senior postdoc at Department of Informatics, King’s College London, The UK and a postdoc at TU Delft, The Netherlands.
My research lies at the intersection of the Semantic Web, Data Protection, AI and Human-Computer Interaction domains and explores how ontologies and knowledge graphs can be utilised to support FAIR, responsible, privacy-preserving and legally (e.g. GDPR) compliant data management as a backbone for trustworthy AI. I completed my PhD at the Semantic Technology Institute (STI) Innsbruck, Department of Computer Science, University of Innsbruck, Austria. My doctoral research focused on supporting machines and humans in making sense of informed consent, for sensor data sharing in smart cities, with knowledge graphs. In collaboration with colleagues, our research on automated GDPR compliance verification based on semantically modelled informed consent was recognised by the OECD Policy Division and the German Federal Ministry of Labour and Social Affairs as a significant contribution to the field of Accountable AI.
In 2023, I joined the European Data Protection Board’s pool of experts in new technologies, privacy and data protection and in 2024 I was awarded a postdoc fellowship in Safety and Security in AI by the German Academic Exchange Service (DAAD). Prior to my PhD, I received an MSc in Advanced Computer Science from Cardiff University, The United Kingdom, where I also completed my BSc in Computer Science and Visual Computing. Some of the projects that I have been involved in include smashHit H2020, CampaNeo FFG, RePlanIT, CIRPASS-2, FAIR-PV, MLCommons.
LATEST NEWS
02/10/2026 – Looking forward to speaking on a panel about the role of data standards in AI at Manchester’s Data Decoded on October 14th, alongside the Open Data Institute (ODI).
04/09/2026 – I am co-chairing The Alan Turing Institute’s Interest Group on Knowledge Graphs.
11/08/2026 – My lab has 3 papers accepted at the SEMANTiCs 2026 Conference:
“Where Structure Helps: RDF vs. Property Graphs for Explainable Legal GraphRAG” by Vishnu Sankarakumar (MSc student) and myself. We compared how property graphs and knowledge graphs perform when providing legal advice regarding the EU AI Act and if semantics matter (spoiler – yes, they do!). We will present results at the International Workshop on Semantic Architectures for Governance and Explainability.
“Governing Multimodal Knowledge Graphs: Current Challenges and Possible Solutions” by Gabriela Kurteva and myself, has been accepted at the Next Generation Data Governance Workshop. We discuss that managing and governing multimodal knowledge graphs across their lifecycle demands a framework that goes far beyond traditional data governance. [Preprint]
“Investigating Stability and Consensus in LLM Taxonomy Generation” by Alison Yearsley and myself, has been accepted as a short paper in the Posters and Demos track. During her BSc thesis work, Alison investigated how good different LLMs are in generating taxonomies. This work is accepted at the main conference.
14/07/2026 – I gave an interview on the importance of data provenance in AI inbreeding (also known as model collapse, AI cannibalism), which is a degenerative process where AI models are recursively trained on data generated by other AI systems. https://agi.fightersteel.com/synthetic-data-could-accelerate-agi-but-damage-human-goals/
19/06/2026 – I have been invited by UK’s Information Commissioner’s Office (ICO) to participate in their upcoming workshops on responsible AI as an expert in responsible data management, AI governance and decentralisation. Looking forward to contributing and helping shape the future of privacy and AI governance.
I was recently interviewed about the importance of having clear and actionable #dataprovenance when developing frontier AI, particularly when training models using synthetic data leading to AI inbreeding (or #modelcollapse). We also discussed the importance of sharing knowledge across fields when the goal is to develop AI responsibly. 
A Multimodal Knowledge Graph Infrastructure for Digital Twinsof Financial Operational Risks: Mortgages and FX stocks – Short poster paper at ESWC 2026 
