• Built end-to-end MLOps pipeline for churn prediction using AWS SageMaker, EC2, and XGBoost. • Created ETL workflows with AWS Glue and Lambda to preprocess customer data stored in S3. • Optimized model using SageMaker Automatic Model Tuning, improving prediction accuracy by 15%. • Deployed trained model via Flask API backend for real-time inference with sub-second response times. • Implemented CI/CD with GitHub, CodePipeline, and CloudFormation, using Docker on ECR. • Managed model lifecycle with SageMaker Model Registry and Step Functions, monitoring performance via AUC.