Tanviben Patel, D.Sc. (Tanvi)

Mentors: Dr. Aldo Badano (Director) and Dr. Seyed Kahaki (Staff Fellow)

Served as an Research Fellow in the Division of Imaging, Diagnostics, and Software Reliability (DIDSR) at the FDA (CDRH/OSEL), focused on the development of computational tools for digital pathology.

• Designed and implemented DataPath, a modular pipeline for WSI preprocessing with capabilities for stain normalization, tissue registration, patch extraction, and metadata-based stratification.
• Developed and validated multi-scanner WSI registration and stain harmonization algorithms, improving reproducibility and cross-platform model compatibility in digital pathology.
• Built tools for outlier detection and cohort stratification to support unbiased AI development and enhance model generalizability across diverse clinical datasets.
• Integrated robust performance evaluation metrics (ROC, AUC, sensitivity, specificity) into assessment workflows, contributing to regulatory-grade validation of AI/ML models for SaMD.

Mentor: Dr. Benjamin Solomon, MD – Clinical Director and Principal Investigator, NHGRI

Completed my doctoral dissertation with a focus on applying deep learning and generative AI for phenotypic analysis in rare disease diagnosis and medical imaging.

• Developed deep learning and generative AI pipelines (StyleGAN2-ADA, HyperStyle, CNNs, SVMs) for facial phenotype analysis to improve diagnostic accuracy in rare genetic conditions.
• Designed expression-invariant classifiers, applied Otsu’s method for image preprocessing, and contributed to explainable AI workflows to enhance clinical interpretability.
• Collaborated in a translational genomics environment and presented findings at NIH as part of a dissertation bridging computer science and genomic medicine.

Advisor: Dr. Hoda El-Sayed, PhD, Professor

• Conducted classification of blood cell types using deep learning models on hematology image datasets, applying ensemble-based methods to improve diagnostic accuracy.
• Developed and evaluated AI pipelines for medical image analysis in hematology, focusing on model robustness, data preprocessing, and performance benchmarking.
• Explored comparative modeling techniques for cancer detection in CT and blood smear images, integrating multi-class classification with high-performance computing tools.

Mentor: Dr. Sharad Sharma, PhD, Professor

• Built immersive Virtual Reality (VR) and Augmented Reality (AR) applications using Unity 3D, with deployment on Android and iOS platforms.

Mentor: Dr. Benjamin Harvey, PhD

• Analyzed COVID-19 Twitter data using Apache Spark and machine learning (K-Means, PCA) to classify misinformation, contributing to the development of a scalable pipeline for social media data analysis.

• Developed full-stack web applications using PHP, JavaScript, HTML/CSS, and MVC frameworks, designing both client-side interfaces and server-side architecture to ensure responsive, cross-platform functionality.
• Built and maintained secure, scalable databases and APIs, collaborated with cross-functional teams to implement user-focused solutions, and optimized application performance through testing, debugging, and mobile responsiveness.

• Designed and maintained responsive front-end interfaces using HTML, CSS, and Bootstrap, integrated data from back-end services, and optimized website performance through thorough testing and debugging.