Maker Profile · Project 01
KidneyGuard
Bringing together hardware, AI and healthcare to explore accessible early kidney-disease risk screening.
What I built
A multi-phase project combining a urine/blood biomarker sample-collection setup with an AI diagnostic layer designed to interpret standard kidney blood-test results and flag early-stage chronic kidney disease risk.
Technical work
Seven machine-learning models were trained and benchmarked on a 399-patient dataset, including logistic regression, random forest and XGBoost. XGBoost achieved 98% accuracy, 1.00 precision and 1.00 ROC-AUC in the project benchmark.
Learning
The project taught me how an open-ended technical project can change direction as mentorship and goals evolve, and the importance of disciplined documentation and planning.