Premtim Sahitaj

Premtim Sahitaj is a researcher and doctoral student affiliated with the Q&U Lab of the XplaiNLP group, focusing on computational approaches to fake news detection under the BMBF-funded news-polygraph project. He holds a B.Sc. in Business Information Systems and an M.Sc. in Data Science from the University of Trier, where he previously worked as a research assistant on the ReCAP project (DFG priority program RATIO) addressing natural language processing in argumentation domains.

His research interests span knowledge-based reasoning (multimodal information extraction, RAG with LLMs, data synthesis), language modeling (multimodal LLM architectures, pre-training optimization), and applications in low-resource languages (e.g., Albanian). He actively supervises student theses and advanced projects through the Q&U Lab.

Prior work includes integrating background knowledge into LLMs' argumentation processes and evaluating information representation strategies. His doctoral research aims to develop computational frameworks supporting investigative journalism through fake news detection platforms.