Projects + Research
Currently, I work on small language models and LLM fairness at Tilburg University. Publications and citations are on Google Scholar.
Research
Master's Thesis: Thematic Bias in Book Recommender Systems
Explored thematic bias in book recommender systems and its impact on users.
Read the thesis or read the shorter version here.
OSAI Index
Contributed to a community-driven resource on model openness, licensing, and EU regulation for generative AI systems.
Projects
An AI Story Blueprint Generator
Built an end-to-end RAG pipeline that turns book metadata into creative story blueprints using Groq LLM and Qdrant.
Deployed a modular Streamlit app with a responsive, accessible interface.
An Open-Source Conversational AI Agent
Built a modular conversational AI pipeline with speech recognition, dialogue management, and text-to-speech using Rhasspy.
Integrated APIs for external data and a T5 paraphrasing model to improve response fluency.
REFLECT: Self-Regulation and Mental Contrasting
Built a mobile application to help students set goals, reflect weekly, and receive AI-driven feedback.
Inspired by the Comenius project on self-regulated learning.
Beyond Medical Records: Endometriosis Storytelling on Reddit
Analyzed Reddit discussions using emotion classification and topic modeling to surface themes of struggles, pain, and surgery.
Validated themes with coherence analysis and compared results to existing research.