I develop transparent AI systems that process legal documents, integrate traceable domain knowledge such as EU legal principles, and produce human-readable explanations grounded in that knowledge.

Developed privacy-preserving, LLM-based solutions that improved user-assistance systems for industrial use cases.

Conducted research on rumour verification and reasoning in LLMs, contributing to two papers.

Developed methods for ranking and matching textual sources in AI-on-demand services, with emphasis on maintaining changing source collections.

Developed NLP methods for large-scale heterogeneous document matching by combining textual knowledge, researcher profiles, user requests, and ontology-based knowledge graphs.

Developed a ranking system for scientific literature based on Argument Mining features.

Developed argumentative dialogue agents for discussing scientific literature and studied how arguments in research papers can improve agent responses.

Developed anomaly detection solutions based on forecasting neural networks and encoder-decoder architectures.