At Icahn School of Medicine at Mount Sinai, I led full-stack development of Harmonizome 3.0, a public gene and protein data portal serving 250k+ annual users. I added dataset-overlap analysis, an AI chatbot, UMAP visualizations, and knowledge graph exports.
I expanded the Harmonizome knowledgebase by 33% with Python and pandas pipelines that integrated 50+ datasets into 30M+ gene-attribute associations.
I operate and maintain the ARCHS4 cloud data-processing pipeline and website on AWS, which aligns and serves 2M+ genomic sequencing samples. I also built LLM-based report-generation workflows that compare gene-set overlaps and produce cited reports with figures.
My work includes classifying retinal OCT images with a ResNet CNN and developing software for a CPR guidance device. I’ve presented technical work at NIH and NCI consortium meetings and annually mentored 2–3 lab trainees.
