Set up the source PostgreSQL database with an outbox table, then run Debezium and an S3 sink connector locally with Docker Compose.

Create a development environment for AWS Glue 3.0 and later by building a custom Docker image, because AWS publishes no image for those versions.

Schedule an AWS Lambda function more often than once a minute, using Amazon SQS to get past the one invocation per minute limit of EventBridge rules.

Build development environments for AWS Glue 1.0 and 2.0 with the published Docker image and the Visual Studio Code Remote Containers extension.

Run frequent Apache Airflow tasks on AWS Lambda instead of Celery workers or ECS, with a custom operator modelled on the ECS operator.

Develop and test AWS applications against LocalStack, a mocking framework that runs AWS services on a local machine, demonstrated with a web service.

Host the web application that calls an R machine learning model on Amazon S3, so both the front end and the Lambda backend run without a server.

Expose an R machine learning model packaged in AWS Lambda through Amazon API Gateway, giving the model a callable HTTP endpoint on AWS.

Deploy an R machine learning model to AWS Lambda and call it without managing a server.

Package a logistic regression model built in R so it can run on AWS Lambda, after testing it locally.