Owned the end-to-end design and deployment of the AML Compliance eFiling module in Java Spring Boot, building a dynamic validation framework with annotations and AOP that streamlined complex compliance workflows and reduced maintenance overhead.
•
Optimized large-scale databases (Yugabyte, ClickHouse, MySQL) powering fraud detection pipelines at 500+ QPS, improving query latency and reliability under production load.
•
Implemented dynamic scaling strategies for Spark and Flink clusters, reducing compute costs by ~30% while maintaining high throughput for UML batch pipelines.
•
Enhanced observability by integrating OpenTelemetry with DeepFlow (eBPF-based), enabling distributed tracing and reducing incident mean-time-to-resolution (MTTR).