Things I've built — open-source frameworks, internal tools, and real-world data engineering solutions.
An object-oriented framework for efficiently handling multiple ETL jobs implementing the Medallion architecture. Built to give teams a structured, reusable way to run and manage complex pipeline workloads with clear bronze → silver → gold separation.
A lightweight library that acts as a clean interface between multiple data stores and data objects. Removes boilerplate from read/write operations, making it easy to swap out underlying stores (Snowflake, Redshift, S3, PostgreSQL) without changing business logic.
Designed and integrated a new Change Data Capture capability into an existing data platform using Debezium and Kafka, enabling real-time ingestion from operational databases. Paired with Great Expectations for automated data quality validation at the point of ingestion.
Designed and delivered a flexible, high-throughput file processing system built on a control tower model. Capable of handling multiple concurrent file formats and processing rules, with full audit trails and error recovery built in from day one.
Established foundational components for Ganit's internal ETL framework and drove its adoption across multiple client projects. The framework standardises how pipelines are built, tested, and deployed — reducing onboarding time for new engineers significantly.
Set up end-to-end observability for data infrastructure using Prometheus and Grafana for pipeline metrics and alerting, and DataHub as the metadata and data lineage catalogue — giving the team full visibility into what's running, what's healthy, and where data comes from.