I'm building AI/ML software systems for TekCommands and Humana, including Neo4j knowledge graph queries, data-quality gates for Python ML pipelines, and AWS workflows across SageMaker, Lambda, and S3.
At the University of Massachusetts Dartmouth, I developed a GNN defense framework that improved biomedical model robustness by 15% under adversarial attack. My first-author IEEE MILCOM paper demonstrated 22% cross-model generalization gains through transfer learning across graph architectures, and I optimized 3D diffusion models to reduce compute costs by roughly 35%.
Previously at Accenture, I built Python backend APIs, optimized SQL procedures and data access patterns, and improved CI/CD pipelines for enterprise applications.
