TimduelStats/MLB-HomeRun-Predictor — reverse-engineered prompt

Build me a Python project that predicts which MLB players are most likely to hit a home run in upcoming games. I want it to use the existing trained XGBoost model and make predictions from batter stats, pitcher matchups, and recent performance data.

Set it up so the prediction job can run in AWS Lambda inside Docker, and make it easy to deploy with AWS CDK. The Lambda should load the model, take in the needed game or player data, produce clear prediction results, and save the output to Amazon S3 so I can review it later.

Please make the code clean and easy to understand, with simple setup instructions for running locally and deploying to AWS. If any details are missing, make sensible placeholders and explain what data format is expected. Look up current AWS docs online if you need to.