At Optum (UnitedHealth Group), I architected medallion data architecture on Azure Databricks for healthcare clickstream data and built ETL pipelines across Snowflake, Apache Airflow, and Azure Databricks. My metadata-driven framework generated Snowflake procedures and Airflow DAGs from YAML configurations, reducing pipeline development time from 40 hours to under 5 minutes.
I also built an AI-driven failure diagnosis engine that analyzes incident logs, generates DDL fixes, and performs root-cause analysis, eliminating 90% of on-call interventions and cutting MTTR by 80%. My work includes healthcare claims fraud signal agents associated with $2M in recovered or prevented losses.

