AWS Glue — serverless ETL in plain words

AWS Glue is a managed service that runs Spark or Python ETL jobs on demand. You hand it a script, point it at your data sources, and it spins up compute, runs the job, and shuts down — you only pay for the seconds it ran.

It also includes a Data Catalog that tracks the schema of every dataset across S3, RDS, and Redshift. This catalog feeds Athena (SQL on S3), Lake Formation (governance), and Redshift Spectrum. For a junior DE, Glue is the easiest way to run real pipelines on AWS without managing infrastructure.