Experience

Developed teacher–student GAN architectures for liver MRI synthesis and segmentation enhancement, working with 6,000+ MRI slices. The approach improved DICE coefficient by 23% compared to baseline segmentation methods, enabling better liver lesion detection and characterization in Eovist-enhanced MRI scans.

Conducting research in machine learning and healthcare applications, working with multiple advisors on diverse projects spanning representation learning, fraud detection, and clinical NLP.

Built production-grade LLM and regex-based pipelines to improve error messaging in enterprise workflow automation systems. Reduced customer wait times by 89% and achieved 99% structural match accuracy across production logs, directly improving system reliability and user experience.

Founded Synapse, an AI-powered research discovery platform that converts real-world science and technology news into structured research questions and relevant academic papers. Built an end-to-end Python pipeline mapping 500+ news queries into arXiv search tasks using LLMs, embeddings, and heuristic ranking over 50k+ papers, enabling scalable discovery of emerging research topics.