Home › Key concepts › Medallion architecture (bronze, silver, gold)

Key concept · Crawl (beginner) · the same on every platform

Medallion architecture (bronze, silver, gold)

Clean data in three layers so each step is simple and you can always go back.

What it is

Bronze holds raw data exactly as it arrived. Silver holds cleaned, de-duplicated, typed data with one row per real thing. Gold holds business-ready tables shaped for a use: a sales mart, a feature table, a metric. Each layer is built only from the one before.

Full guide: Medallion architecture, drawn right, with every layer, one order followed through them all, the rules, the mistakes and the products on each platform.
Sourcesapps, files,APIsBronzeraw, as itarrivedSilverclean, typed,one row per thingGoldshaped fora useUsesdashboards,ML, AIeach layer is built only from the one before, so you can always rebuild
Medallion architecture: raw, clean, ready. Fix a step and rebuild from the layer below.

Why it matters

When something is wrong you can see in which layer it went wrong, fix that step and rebuild from the layer below.

An example

Bronze: raw order JSON. Silver: one clean row per order with proper types. Gold: daily sales by store.

The common mistake

Letting dashboards read bronze directly, so every report re-does the cleaning differently.

Where you meet it on each platform

The idea stays the same; only the product name changes.

JobAWSAzureGoogle CloudDatabricksOpen source
Open table formatfiles that behave like a database table, with historyApache Iceberg (S3 Tables)Delta Lake in OneLakeIceberg on BigLakeDelta LakeDelta Lake or Iceberg
SQL modellingturn raw tables into clean, tested, documented onesdbt on Redshift or Athenadbt on FabricDataform or dbt on BigQueryLakeflow Declarative Pipelines or dbtdbt with DuckDB

Related ideas

Where it shows up

Part of the Crawl level on these paths: Data Engineering (shaping data) · ML Engineering (features and data).