At Hometrack, I owned an end-to-end Databricks ETL pipeline for government energy performance certificate data, replacing a legacy SQL Server Integration Services process. It saved about two hours of manual effort each week and addressed customer complaints about missing data.
I also built a PySpark schema check that fails in minutes instead of late in a production build, and Python automation that stopped around 10 monthly load failures. Previously, at SpecFin Capital, I analysed UK mortgage data for underwriting and credit risk, contributing to a 30% reduction in financing costs.

