Machine Learning Researcher
I’m a Final Year Computer Science PhD student under the supervision of Prof. Raul Santos-Rodriguez in Interactive Artificial Intelligence.
My research spans explainable AI, cost-sensitive learning and human image perception. I have academic and industrial software development experience. I am lead developer of the IQM-Vis toolbox.
I also work part time over at GreenPixie as the lead data engineering following an internship during my third year of PhD. At GreenPixie I helpi to report and reduce enterprises’ cloud carbon footprint by creating hardware energy methodologies and implementing highly optimised production code to process large (TBs) of cloud cost and usage reports (CURs) using polars.
See my Google Scholar for an up to date list of publications.
Learning to classify with imbalanced data, how can we optimally set the bias term of a classifier to take into account the uncertainty of two classes with respect to their imbalance qualities. Paper
When can post-hoc explanations of models become unfaithful, and how reconciling a mismatch between training and evaluation distributions can increase faithfulness. Paper, Blog Post
Evaluating how the human visual system’s understand of the natural world through images can be utilised for machine learning and vision tasks. My IQM-Vis toolbox can be used for this, and was featured in the ICCV’23 demos session and published in the September 2025 issue of the SoftwareX Journal.
On going research into analysing data before learning to help guide inductive bias choices.