At AMAN Holding, I re-architected E-Payment reconciliation reporting, cutting daily execution time from 2–3 minutes to under one second. I also maintain production pipelines and built monitoring and data-quality checks to surface issues before they reach downstream consumers.
I built lakehouse infrastructure with Hadoop HDFS and Apache Iceberg, integrating Oracle, PostgreSQL, and SQL Server sources. I also implemented a PySpark commission calculation engine that replaced a manual Excel process and supports SCD Type 2 rules.
At Zid, I designed a Medallion architecture on Amazon Redshift and built more than 150 dbt models. I helped transition the warehouse from ERD to dimensional modeling, reducing Metabase dashboard load times by up to 80%.
