srushti User

At Amazon, I built an event-driven automation service for seller grace-period validation using Java, AWS Lambda, SQS FIFO, SNS, and internal services. It reduced investigator effort by 40%, eliminated 366+ manual hours per month across 7,300 weekly cases, and delivered $40K in annual operational savings.

I also developed a metric recalculation framework for more accurate post-reinstatement evaluations, migrated production MAWS and NAWS pipelines for multi-region fault tolerance, and upgraded production services to JDK 17.

At KPIT, I created predictive-maintenance workflows for fuel-cell Remaining Useful Life, training and evaluating seven regression models. I also built an Isolation Forest and SHAP anomaly-detection framework that analyzed 49,512 operational data points and identified 1,258 anomalies with interpretable diagnostic reports.

My projects include an AI DevOps Agent for cloud reliability automation and an ecommerce BI platform that processes 151K+ records through validated Python and SQL ETL pipelines. I enjoy building reliable backend, cloud, and data systems with Java, Python, AWS, SQL, and distributed-systems fundamentals.