Shubham Ovhal

I'm completing an MSc in Data Science and Artificial Intelligence at Newcastle University, where I build and evaluate AI systems with rigorous statistical methods rather than relying on headline metrics alone.

For my cardiac MRI vision-language dissertation, I implemented a debiased Centered Kernel Alignment metric from scratch after identifying a flaw in the standard estimator. I benchmarked four architectures using five-fold cross-validation and patient-level DeLong's tests, reporting a genuine null result when no pairwise differences were statistically significant.

I've also built segmentation, classification, and population-health analytics pipelines using Python, MONAI, Power BI, Alteryx, and SQL — including a 3D cardiac segmentation model deployed live on HuggingFace, taken from training through to a publicly usable application. At Newcastle University, I translated athlete-performance data into live dashboards and practical recommendations for coaching staff, while consulting projects strengthened my ability to turn analysis into clear stakeholder strategies.