At Ominimo Insurance, I built and unit-tested PySpark ingestion and transformation jobs to onboard a new market onto a medallion-architecture data platform. I also designed a 100+ feature pipeline and XGBoost models that predicted competitor prices with 13–20% MAPE.
I own a production reporting pipeline delivered semi-monthly to external partners and the reinsurer, reviewing transformation logic against business rules and fixing data-integrity defects before each release.
At Robert Bosch, I re-implemented 15+ KNIME workflows in Python and SQL, with output-equivalence checks as regression tests. In my projects, I built data and backend systems, including a math-solving service and an ingestion pipeline.

