Dario Di Palma

Email / CV / Scholar / Linkedin / Github

Hi, I’m Dario Di Palma, a Ph.D. in Artificial Intelligence specializing in Large Language Models and Recommender Systems.

I earned my PhD at Politecnico di Bari, Italy, where I conducted research on LLM interpretability, recommendation methodologies, and agentic AI systems. I was advised by Prof. Fedelucio Narducci, and Prof. Vito Walter Anelli, within the research group led by Prof. Tommaso Di Noia.

Research

I am interested in Large Language Models, Recommender Systems, Representation Learning, and Agentic AI. Most of my PhD research has focused on developing methodologies to interpret LLMs, integrating them into recommendation pipelines, and leveraging them effectively for recommendation tasks. During my PhD, I received the Best Short Paper Award at SIGIR 2025 for uncovering the memorization of recommendation datasets during LLM training.

I developed an interpretability framework that reduced memory usage by 50 percent and accelerated inference time for sentiment analysis with LLaMA. I also studied Automatic Prompt Engineering techniques to speed up the discovery of effective prompts for specific tasks. I collaborated with IBM Research on Text-to-SQL, proposing a novel test-time inference method. In addition, I investigated the factuality of LLMs, analyzing their ability to self-detect when they produce incorrect or misleading statements.

Supervisor, Mentorship and Dissemination

I am open to supervision and collaboration. If you are interested in conducting research on LLMs, Recommender Systems, or Agentic AI, please send an email using the button below with the subject line “[Research Collaboration – YOUR LAST NAME]” and include the required materials.

📩 Email Me

Please attach:

• Your CV
• A motivation letter describing your research interests and explaining why you would like to work with me