Selected academic, clinical, and AI projects at the intersection of medical physics, radiation medicine, and healthcare technology.
First-author benchmark study testing whether patient-similarity graph neural networks improve multi-outcome prostate-cancer survival prediction compared with well-tuned classical models.
Quantitative analysis of international radiotherapy audit data to assess quality-improvement maturity, using quality indicators, scoring methodology, and benchmarking of radiotherapy centres.
Python-based software for independent dose evaluation and verification in OSLD/RPLD dosimetry audits, providing a secondary verification workflow for audit dose calculations.
Collaborative imaging project simulating low-dose PET from clinical list-mode data and restoring full-dose image quality with a Vision Transformer, toward lower dose or shorter acquisition.
Healthcare-AI industry work building deep-learning pipelines for dental image segmentation, classification, enhancement, and image generation to support diagnostic workflows.
Personal project using a frozen pathology foundation-model encoder and a linear probe for tumor-vs-normal classification on PatchCamelyon histopathology patches.
Personal project building a multi-step agentic workflow that retrieves biomedical papers, extracts key findings through tool-calling, and answers structured research questions.