Research roles across clinical AI, biostatistics, causal inference, and computational modeling.

Project Co-Investigator & Visiting Scholar

Reliability and Failure Analysis of Multimodal Clinical AI

MIT Laboratory for Computational Physiology · MIT Critical Data

I evaluate ensembles of diagnostic-imaging models and quantify how disagreement accumulates across images and tasks. I then link those patterns to clinical context in MIMIC-CXR and MIMIC-IV. Calibration, uncertainty, threshold sensitivity, temporal linkage, and subgroup performance are part of the primary analysis.

Lead Data Scientist & Biostatistician

Surgical Decision Support for Epilepsy Outcomes

Duke Comprehensive Epilepsy Center & Department of Neurology

I lead the statistical design and reproducible analysis of whether the SEEG-derived 5-SENSE focality score is associated with postoperative seizure freedom. The work evaluates discrimination, predictive values, ROC/AUC with bootstrap confidence intervals, calibration, proposed thresholds, and failure patterns across resection, LITT, and neuromodulation cohorts.

Postdoctoral Associate

ML-Integrated Multiscale Modeling for Biomedical Systems

Duke University · Department of Biostatistics & Bioinformatics

I developed reproducible workflows that connect mechanistic simulators with statistical learning and neural models. The work includes gradient-boosted trees with Bayesian optimization, Gaussian-process surrogates, physics-informed neural networks, and cross-species alignment by optimal transport with relaxed marginals.

Joint NIH-funded Duke–Weill Cornell Medicine congenital-CMV collaboration · R01 AI173333.

NSF NRT Research Fellow in AI & Machine Learning

Causal-Network Inference & Interpretable ML

Bridges to Digital Health · Award #2125872 · West Virginia University

I designed a causal-network workflow for high-dimensional biomedical measurements. It combined probabilistic graphical models, constraint- and score-based structure learning, multiple-testing-aware conditional-independence analysis, and bootstrap assessment of edge stability. This work developed into BaMANI.

Machine Learning Researcher

Generative & Causal Machine Learning for High-Dimensional Biology

WVU Cancer Institute · Machine Learning in Systems Biology Lab

I developed variational autoencoders and Wasserstein GANs for representation learning and synthetic-data generation, Bayesian structure-learning workflows for causal discovery, and deep-learning methods for biological signal decomposition. The work supported studies of CCN4/WISP1 and tumor-immune network organization.

Research Scholar & Project Lead

Computational Applied Mathematics

Shiraz University of Technology

I developed numerical algorithms for nonlinear dynamical systems using spectral, finite-element, and analytical methods. This work produced ten peer-reviewed articles in computational and applied mathematics and included mentorship of three M.Sc. students.