Thukaraka Pakeerathan

Thukaraka’s PhD research investigates how Large Language Models (LLMs) can be used to build recommender systems that are not only accurate, but also fair and diverse. Her work examines how prompt design and generation settings influence which items and providers receive exposure, with particular attention to factors such as popularity, gender, location, and language. The research also explores automated prompt optimisation as a more reliable alternative to manual prompt tuning, with the aim of developing recommendation systems that balance user utility with fairer and more diverse outcomes.

Thukaraka Pakeerathan is supervised by Prof. Jeffrey Chan, Dr. Ziqi Xu and Dr. Danula Hettiachchi.