Keerthana Reddy

Aws Data Engineer @Fifth Third Bank

Chicago, IL, US

Designed and implemented scalable ETL/ELT pipelines using AWS Glue, Databricks, and Python to process structured and semi-structured AML and transaction datasets. Built end-to-end data ingestion and transformation frameworks to migrate high-volume financial data from on-prem sources into AWS S3, Redshift, and Snowflake. Designed and optimized data models and feature stores to support AI/ML-driven fraud detection and compliance monitoring use cases. Partnered with data scientists to prepare, label, and transform training datasets used for AI/ML models focused on suspicious activity detection and risk scoring. Integrated AWS SageMaker and Databricks MLflow pipelines to manage model training, versioning, and deployment of fraud detection models in production. Worked on feature engineering, data validation, and drift detection processes to maintain model accuracy and ensure high-quality data input for predictive analytics. Developed real-time data pipelines using Kafka/MSK and Spark Structured Streaming to feed AI/ML inference workflows for instant fraud alerts. Automated retraining workflows to refresh features, update datasets, and trigger model evaluations through Airflow and Lambda. Collaborated with business and compliance teams to evaluate AI model outputs, monitor anomalies, and refine rule-based AI scoring thresholds. Implemented observability and quality frameworks in Glue and CloudWatch to track schema drift, null anomalies, and ML performance metrics.