At Axis Bank, I build near-real-time lending data solutions that cut loan application visibility to under 10 minutes and improve operational efficiency. I design ETL and streaming pipelines with Python, SQL, PySpark, Apache Kafka, Informatica, and Apache Spark, including a migration from Hive to Apache Iceberg that lowered query latency by approximately 40%.
I've also deployed an XGBoost mortgage foreclosure propensity model validated on Out-of-Time data with 0.78 AUC-ROC and 72% recall. My work spans data modeling, cloud pipelines, CI/CD, and stakeholder collaboration across credit, lending, and risk teams.
