4 About Medallion Architecture
Medallion architecture is a layered design pattern for organizing and processing data in a data lakehouse environment.
- Trusted, high-quality data
- Traceability and auditability
- Modular, efficient pipelines
AI Data Platform Workbench provides you with the tools to establish your bronze, silver, and gold stages and automate cleaning, validating, and enriching your data through each medallion layer.
The bronze layer stores raw data ingested from your source systems, preserving the data in its original form. It acts as your single source of truth and supports traceability and auditability. Data is pulled directly into AI Data Platform, preserving the original data structure and supporting fast, scalable retrieval for downstream processing. Data is then ingested into the bronze layer and is ready for processing.
Data from the bronze layer is processed and sent to the silver layer. Workflows, agents, and other automation help you clean, remove duplicates, and transform your data so it is analytics-ready. Data in the silver layer is often joined across sources to provide standardized, consistent information.
In the silver layer, data analysts manage and develop the cleaned data to prepare it for consumption in data dashboards, advanced analytics, and machine learning models. Data prepared this way is stored in the gold layer, and data in the gold layer is treated as the most trustworthy, and best suited for training models and developing data visualizations, dashboards, and more.