At RFC, I built a commercial decision-support system from scratch on a Microsoft Fabric Lakehouse, replacing a legacy Data Warehouse. I designed incremental PySpark pipelines and used Delta Lake change data feeds to propagate updates through the medallion layers.
I also implemented automatic pause and resume for Fabric capacity, keeping a Power BI report available while reducing idle compute costs. The system included models for opportunity win-probability and customer segmentation, with results displayed in Power BI.

