I am a multidisciplinary researcher with a background in cognitive science, machine learning, and NLP. Currently, I am a Research Fellow (Scientific Project Officer) at the European Commission Joint Research Centre (JRC), in the digital health unit. My work focuses on explainability and multimodality. In particular, I am interested in integrating graph-based representations and biological domain knowledge within LLMs, and using explainability and mechanistic interpretability to understand the behaviour of multimodal biomedical models and mitigate implicit biases that influence their diagnostic decision-making processes

Before joining the JRC, I was a postdoctoral researcher in the Artificial Intelligence Research Group at the University of Sussex, where I also earned my PhD in Informatics. My thesis examined how LLMs and other machine learning models learn to handle linguistic abilities such as compositionality, generalisation, and inference.

Occasionally, I work on problems related to dream analysis, such as adopting LLMs to annotate and analyse dream content for research purposes.

News

• January 2026: I will be at the MIT DreemXEngineering Symposium to present a work on multimodal xAI for dream detection in EEG signal.
• September 2025: We will be presenting 4 full papers at the 11th International Conference on Machine Learning, Optimization, and Data Science(LOD 2025) with subject varying from text-to-graph to dream analysis (see publication page for titles).
• July 2025: We will present our paper on Task-level XAI for biological sex biases in the medical domain at the Gender Bias in NLP (GeBNLP) workshop at ACL 2025.
• July 2025: Our work on conformal prediction for nodule detection has been accepted for publication at COPA 25 conference.
• June 2025: Our work Dreams are more “predictable” than you think has been accepted for publication.