Add flows which train a model and run inference from the trained model

[daniel-prefect]

Related to https://linear.app/prefect/issue/DOC-118/create-an-author-tutorial-for-model-training

Merge this before PrefectHQ/prefect#16632

Here's the full workflow that these flows enable:

1. Upload training data to an S3 bucket, this triggers a webhook event that triggers a Prefect automation
2. The Prefect automation runs a training flow (model_training.py), which uses Sagemaker to fit an XGBoost model to the training data, and then uploads the fitted model to a second S3 bucket.
3. When a model is uploaded to the second S3 bucket, that bucket triggers a webhook event that triggers a second Prefect automation.
4. The second Prefect automation runs an inference flow (model_inference.py), which loads the fitted model from S3 and generates predictions for a small set of test values.

NOTE: The code for using XGBoost with the Sagemaker Python SDK is based on these docs: https://sagemaker.readthedocs.io/en/stable/frameworks/xgboost/using_xgboost.html