At Stripe, I engineered 80+ real-time and batch features for fraud, churn, and payment-routing models, reducing feature turnaround for data scientists by 45%. I also built real-time pipelines processing 2M+ transactions per second and delivering sub-second signals to ML models.
I integrated MLflow for experiment tracking and model versioning across Snowflake and AWS, reducing model-to-production cycle time by 30%. I also designed a Delta Lake architecture that cut batch costs by 28%.
At Fractal (Analytics Consulting), I productionized feature engineering pipelines for predictive forecasting models, helping reduce client supply chain stockouts by 25%. I also delivered Power BI dashboards and engineered ETL/ELT pipelines for analytics.
