I am a Research Assistant working with Dr Paul Kirk as part of the Precision Medicine and Inference for Complex Outcomes (PREM) theme at the Biostatistics Unit. My current work is focused on applying Bayesian outcome-guided clustering to identify clinically relevant subgroups in patients with Alcoholic Hepatitis with the aim of improving patient outcomes. This work is forms part of the larger research effort of the MIMAH consortium.

I have a background in biological sciences, with a masters in Theoretical Systems Biology and Bioinformatics from Imperial College London. After my masters, I stayed on for my PhD as a Wellcome Trust Scholar.  Here, I worked with Professor Michael Stumpf on learning cell-fate decisions boundaries from single-cell transcriptomic data with interpretable machine learning methods. As part of this research, we also developed Bayesian tree-based ensembles capable of incorporating prior information to improve predictive performance and embedded feature selection in sparse contexts.

More generally, I am interested in feature selection problems, supervised and unsupervised learning, and leveraging Bayesian statistics to improve human health outcomes.

Selected papers

• “Decision tree models and cell fate choice”, I. Croydon Veleslavov and M. Stumpf, arXiv preprint 10.1101/2020.12.19.423629
• "Repeated Decision Stumping Distils Simple Rules from Complex Data", I. Croydon Veleslavov and M. Stumpf, arXiv preprint 10.1101/2020.09.08.288662
• “Data Study Group Final Report: Greenvest Solutions - Forecasting Wind Energy Production Using Satellite Data”, T Gupta et al., https://www.turing.ac.uk/research/publications/data-study-group-final-report-greenvest-solutions