pavannagula/Fraud-Detection-E2E-ML-project-using-MLflow ? reverse-engineered prompt

Build me an end to end credit card fraud detection app that can tell whether a transaction looks legit or fraudulent.

I want the whole workflow in place, from pulling the data in, checking that it matches the expected columns, cleaning and transforming it, training a model, and saving the best one for later use. Please use the Kaggle fraud dataset style setup, and make the training pipeline reproducible with experiment tracking so I can compare runs and see which model performs best.

Then wrap it in a simple web app where someone can enter transaction details and get a fraud prediction back right away. It should be easy to run locally, but also ready for deployment with Docker and AWS, including EC2 and ECR, and a basic CI/CD flow through GitHub Actions if that fits. If you need to look up current docs online for any of the tools, go ahead.