Demonstrating how neural networks manage to avoid overfitting through "sloppiness"—a unique pattern where models exhibit varying parameter sensitivities, effectively controlling generalization despite extensive parameters.Suggesting more streamlined and efficient AI development practices by uncovering common learning patterns across various deep learning models.An Independent cutting-edge graduate-level project under supervision of Dr. Pratik Chaudhari receiving support from major grants and awards, including the NSF CAREER award and partnerships with industry leaders like Amazon and Intel.

An independent graduate-level project of developing mathematical models to explore economics model of evolutionary game theory and switching costs. Under supervision of Dr. Akcay and Dr. Plotkin

Build and trained extensive machine learning models to understand root of absenteeism in Dallas Area Rapid Transit (DART) company's worker. Worked with Python and SQL to analyze DART's datasets.