At Gupy, I’m leading the migration of our data architecture to Databricks, including the pipeline approach, standards, and CI/CD workflow. I migrated 28 billion records across 119 tables into Databricks in under two hours and added validation and data quality checks.
I also operate Gupy’s Apache Iceberg lakehouse and built its maintenance pipeline in Java, which contributed to a nearly 30% reduction in monthly AWS spend. I led the Airflow 2 to Airflow 3 migration, cutting a near real-time dbt pipeline from 30 minutes to 9.
For analytics, I led the move from Cube Cloud to self-hosted Cube.js on StarRocks, bringing dashboard query latency from up to 60 seconds to around 400 ms. Together with my team, I built MCP servers that power the AI chat inside the product and support internal analysts.
Earlier at Gupy, I worked across full-stack application and data products and was tech lead of a six-person team. At UEMA, I rebuilt the university entrance exam registration system, which drove a 150% revenue increase in 2018 versus the legacy system. My projects also include open-source observability tools and an Iceberg maintenance control plane.
