Sai Kiran
Reddy Tallapureddy

Engineered a real-time fraud detection pipeline processing 10,000+ transactions per second using Databricks Medallion Architecture. Built streaming ingestion with Kafka, applied PySpark transformations through Delta Live Tables, trained an XGBoost classifier achieving 90%+ accuracy, and served insights via Power BI dashboards with sub-second latency.

Built an end-to-end demand forecasting platform ingesting 20,000+ events/day from IoT sensors and POS systems via Kafka. Applied Prophet and XGBoost ensemble models for multi-horizon forecasting, reducing predicted waste by ~20%. Deployed on Kubernetes with autoscaling, tracked experiments through MLflow, and stored feature sets in Delta Lake for reproducibility.

Designed and deployed an autonomous Kubernetes reliability platform with self-healing capabilities. Integrated Prometheus for metrics collection, Grafana for visualization, Loki for log aggregation, and Alertmanager for intelligent alerting. Built Python-based remediation controllers that automatically detect and resolve pod failures, resource bottlenecks, and SLO violations with zero human intervention.