Research & Publications

100+ Peer-Reviewed Publications

A prolific research career spanning computational modeling, medical imaging, machine learning, and clinical decision support — with work published in The Lancet, JACC, and leading engineering journals.

Google Scholar Metrics

View Profile
1,128
Total Citations
13
h-index
27
i10-index
69+
Publications
AI & Machine Learning

AI/ML Research Publications

Dedicated research in machine learning, deep learning, transformers, and applied AI — from medical imaging to breath-based diagnostics.

COVID-19 biomarkers in human breath using Proton Transfer Reaction Time-of-Flight Mass Spectrometry

EClinicalMedicine (The Lancet)2021

Developed breath-based COVID-19 detection using ML applied to PTR-ToF-MS data from breath samples collected via silicon microreactors. Identified volatile organic compound biomarkers for non-invasive screening.

ML Classification + PTR-ToF-MS
56
citations

CTEPH AI Diagnostics — CT Pulmonary Angiogram Detector

Bayer Global Digital Innovation Award2019

Built AI detector for Chronic Thromboembolic Pulmonary Hypertension from CT pulmonary angiograms. Productionized on Google Cloud, submitted for FDA 510(k) as Class 2 medical device. Organized global multi-center clinical trial.

Deep Learning + Medical Imaging + GCP
Featured

Use of Electrocardiograms to Identify Coronary Artery Disease: Cross-Validation of an Artificial Intelligence Model

JACC: Advances (American College of Cardiology)2026

Large-scale validation of ECGio deep learning model on 16,476 patients undergoing coronary angiography. Achieved AUC of 0.914 in cross-validation and 0.924 in external validation for predicting clinically significant CAD from standard 12-lead resting ECGs. High negative predictive value positions ECGio as a potential gatekeeper to reduce unnecessary invasive angiography.

Deep Learning + ECG (ECGio / HEARTio)Read Paper

Classification of Echocardiography Videos Using TimeSformer for Detecting Incipient Heart Failure

AHA Circulation (American Heart Association)2024

Employed TimeSformer video transformer to classify echocardiography videos for detecting incipient heart failure in asymptomatic patients with normal ejection fraction.

TimeSformer Video Transformer

Gradient Boosting Model for Corporate Default

The Journal of Fixed Income2024

Used extreme gradient boosting (XGBoost) to predict 1-year corporate default probabilities from 1M+ monthly estimates. 50 input variables achieving 0.8 correlation with DRISK and StarMine PDs. Effective for bond relative value trading.

XGBoost / Gradient BoostingRead Paper

Convolutional Neural Network based Segmentation of Abdominal Aortic Aneurysms

IEEE Conference on Biomedical Engineering2021

Proposed a 3D UNet CNN architecture for voxel-wise segmentation of AAAs including aorta, aneurysm sac, intra-luminal thrombus, and calcifications from CT images.

3D UNet / CNN

Image Transformers with Regional Attention for Classification of Aneurysm Rupture Risk

SPIE Medical Imaging2022

Applied image transformers with regional attention mechanisms to classify abdominal aortic aneurysm rupture risk without requiring explicit segmentation.

Vision Transformers (ViT)

Classification of Abnormal Cardiac Rhythm from Brief Single-Lead ECG Recordings using Transformer Encoders

IEEE2024

Applied transformer encoder models with transfer learning to classify abnormal cardiac rhythms from brief single-lead ECG recordings.

Transformer Encoder + Transfer Learning

cMRI-BED: A Novel Informatics Framework for Cardiac MRI Biomarker Extraction and Discovery Applied to Pediatric Cardiomyopathy Classification

BioMedical Engineering OnLine2015

Novel informatics framework for automated extraction and discovery of cardiac MRI biomarkers, applied to pediatric cardiomyopathy classification using machine learning methods.

Machine Learning + Cardiac MRI Biomarkers
32
citations

Deep learning model for end-to-end stratification of thrombotic risk in left atrial appendage

Engineering Applications of Artificial Intelligence2023

Efficient end-to-end deep learning model for stratifying thrombotic risk in the left atrial appendage from medical imaging data.

Deep Learning + Medical Imaging

Automatic segmentation of the left atrium from computed tomography angiography images

Annals of Biomedical Engineering2023

Automated segmentation framework for the left atrium from CTA images, enabling cardiac function analysis.

CNN Segmentation
6
citations

3D Segmentation of Abdominal Aortic Aneurysm Walls from CT Angiograms

IEEE International Conference on Bioinformatics and Biomedicine2022

AAA wall segmentation from CT angiography imaging for understanding disease size, location, and geometry. Addresses the challenge of segmenting thin, annular vascular structures.

3D UNet / Deep Learning
3
citations
Top Cited

Highest-Impact Publications

Selected publications ranked by citation count from Google Scholar.

1

Variability of computational fluid dynamics solutions for pressure and flow in a giant aneurysm: the ASME 2012 Summer Bioengineering Conference CFD Challenge

J. Biomechanical Engineering2013
194
citations
2

Real-world variability in the prediction of intracranial aneurysm wall shear stress: the 2015 international aneurysm CFD challenge

Cardiovascular Eng. & Technology2018
138
citations
3

Alk1 controls arterial endothelial cell migration in lumenized vessels

Development2016
137
citations
4

The computational fluid dynamics rupture challenge 2013 — Phase II: variability of hemodynamic simulations in two intracranial aneurysms

J. Biomechanical Engineering2015
125
citations
5

COVID-19 biomarkers in human breath using Proton Transfer Reaction Time-of-Flight Mass Spectrometry

EClinicalMedicine (The Lancet)2021
56
citations
6

Effects of intraluminal thrombus on patient-specific abdominal aortic aneurysm hemodynamics via stereoscopic particle image velocity and computational fluid dynamics modeling

J. Biomechanical Engineering2014
54
citations
7

Physics based image processing and evaluation process of perfusion images from radiology imaging

US Patent 9,179,8812015
47
citations
8

Growth and hemodynamics after early embryonic aortic arch occlusion

Biomechanics & Modeling in Mechanobiology2015
46
citations
9

Characterization of neonatal aortic cannula jet flow regimes for improved cardiopulmonary bypass

J. Biomechanics2013
44
citations
10

Aortic outflow cannula tip design and orientation impacts cerebral perfusion during pediatric cardiopulmonary bypass procedures

Annals of Biomedical Engineering2013
40
citations
Intellectual Property

Patents

7 unique inventions across 8 granted US patents, spanning healthcare AI, neuroscience, computer vision, biomedical engineering, surgical navigation, and medical imaging.

Expertise

Research Areas

Core research domains spanning engineering, medicine, and computer science.

Computational Fluid Dynamics (CFD)Cardiovascular BiomechanicsMedical Image AnalysisMachine Learning / Deep LearningNatural Language ProcessingComputer VisionHealthcare AI / Clinical Decision SupportMass Spectrometry AnalyticsKnowledge GraphsTransformer ArchitecturesSignal ProcessingAgentic AI Systems