What are the three layers of buckets in a data lake, and how does an S3 event trigger an AWS Glue job that cleans data and loads it into the silver layer?

A common bucket‑based data lake architecture uses three logical layers: bronze (raw), silver (cleaned), and gold (curated). Each layer resides in its own S3 bucket or prefix. When a new file lands in the bronze bucket, an S3 event notification (via EventBridge or S3 Event Notifications) triggers an AWS Glue job. The Glue job reads the raw data, applies transformations such as schema inference, type casting, and data cleansing, and writes the cleaned output to the silver bucket. The job can be scheduled or event‑driven, and it typically uses Glue ETL scripts or Spark jobs. This decouples ingestion from processing, allows auditability, and ensures that downstream consumers only access the cleaned silver data.