Nabya Faisal

I've built AI-driven biomedical engineering projects that turn complex clinical data into practical, usable tools. At Riphah International University, I developed an end-to-end Python pipeline using conditional diffusion models to generate synthetic ECG and PPG signals on demand.

I designed a 15-metric evaluation framework using Fréchet Distance, DTW, SSIM, and classifier-based accuracy to validate generated output. I also deployed the solution as a full-stack web application where non-technical users could generate and export signals as CSV and PNG files.

For a brain tumor detection project, I trained a VGG16 transfer-learning model to classify MRI scans across glioma, meningioma, pituitary, and no-tumor categories, achieving 96% classification accuracy. I automated preprocessing, augmentation, training, and evaluation while documenting the methodology and results in a formal report.

During my internship at Islamabad Medical Complex, I reviewed recurring faults in patient monitors, ECG machines, and infusion pumps, supported troubleshooting with senior engineers, and produced structured service records. I enjoy translating biomedical, software, and AI concepts into clear technical documentation that improves quality, traceability, and decision-making.