At Johnson & Johnson, I architected and built iDEE for Databricks, a metadata-driven ETL framework that lets teams onboard a pipeline by authoring a YAML config instead of writing code. It runs pipelines across landing, bronze, silver, and gold layers through a shared engine.
I also led gene-signature discovery on a Johnson & Johnson prostate-cancer clinical-trial dataset. The analysis resolved 13 signatures and included enrichment validation and a reproducible pipeline for clinical association and survival analysis.
For MedTech CCP, I designed and delivered a Python ETL pipeline that replaced Tableau Prep Builder and cut pipeline rebuild time from about two months to 2.5–3 weeks. I also migrated Treasure Data CDP environments to Microsoft Entra with no production downtime.
At Johnson & Johnson – Ethicon, I developed Janytor, a Python Dash application that reduced data processing time from about 30 minutes to 7 seconds per dataset. My earlier graduate research at Boston University included bioinformatics pipelines and interactive tools for genomics, transcriptomics, and structural biology.
