Om Prakash

At JCPenney India, I architected and delivered a Django-based Federated Data Platform that lets non-technical users configure and deploy PySpark pipelines without code, reducing pipeline onboarding time by 70%.

I build enterprise data platforms across AWS, Snowflake, EMR, Airflow, dbt, and DataHub, with Medallion Architecture, PII tagging, lineage, and Great Expectations validation embedded throughout the pipeline lifecycle. I also led the migration from a legacy Cloudera Hadoop cluster to Hive on Amazon EMR and improved critical ETL performance by 50% through query tuning and parallelization.

Previously at Xpressbees, I designed real-time and near-real-time pipelines using Kafka, Debezium, AWS, Snowflake, and Airflow. I integrated machine learning models into production pipelines, improving predictive analytics accuracy by 25%.

At Condé Nast, I maintained and optimized more than 100 production pipelines and led a migration from Luigi to Apache Airflow, improving observability, reliability, alerting, and SLA adherence. I enjoy building reliable, governed data products that make high-quality data easier for teams to use.