At DataMount, I cut the tax rule engine’s end-to-end runtime 73% by refactoring its execution model with multithreading and moving single-node, high-memory workloads to Polars. Following my promotion, I took ownership of the engine’s architecture and delivery, extending it with operator support, navigation trees, and per-partner reporting for a Big Four client on Azure Databricks.
I also delivered production batch ETL pipelines at DataMount using Spark Declarative Pipelines, medallion architecture, and idempotent design. In my projects, I built a Chicago taxi Delta Lake pipeline processing about 5 million rows per month and helped scale COMSATS Plus to more than 7,500 active users.

