Research, teaching, and applied work in statistics and machine learning.
T-DIET: longitudinal dietary patterns and health outcomes
Statistical methods for modelling how dietary patterns change over time and how they relate to health outcomes. Supervised by Assistant Professor Silvia D'Angelo and funded by the Research Ireland Frontiers for the Future Programme.
Large language models and sequential decision-making
Tested how well large language models solve sequential decision-making problems, using the multi-armed bandit problem as the main setting. Compared LLM behaviour against classical bandit algorithms and examined where the models fail when reasoning under uncertainty over many rounds. Supervised by Prof Audrey Durand, Université Laval.
Open course material
Three self-contained course websites: Bayesian statistics and inference, hidden Markov models, and R programming from beginner to expert. See the Courses page for details.
IFRS 9 credit risk modelling
Built and validated IFRS 9 credit risk models for bank clients, covering probability of default (PD), loss given default (LGD), exposure at default (EAD), and macroeconomic variable (MEV) components.
Science communication workshop
Completed the Africa Scientifique workshop on science communication, which covered scientific writing and presenting research to non-specialist audiences.