Data Warehouse product
Lake
Catalog data models can use
Ingest CSVs, database tables, and unstructured archives into bronze. Curate silver views and publish gold datasets so ML, Forecast, and Loop train on clean tables — not spreadsheet exports.
Start with a CSV or a database. No data science team required.
Medallion architecture
Bronze → Silver → Gold, in one catalog
Every dataset lands in bronze, gets curated in silver, and publishes to gold when it's ready for ML, Forecast, or Loop.
Bronze layer
Capabilities
Your Data Warehouse foundation
Ingest from spreadsheets or databases, curate silver layers, and publish gold datasets for modeling.
Bronze ingest
Upload CSV and Excel, connect databases, or ingest unstructured zip archives — all into bronze storage.
Silver curation
Define silver views and filters so teams work from consistent, queryable datasets.
Gold publish
Publish gold artifacts ready for ML, Forecast, and Loop training pipelines.
Model-ready catalog
Browse, preview, and connect published gold datasets to intelligence products in one place.
Use cases
What teams build with Lake
From consolidating messy sources to publishing model-ready gold — common warehouse workflows on Heimdall.
Browse all warehouse and intelligence scenarios on the use cases hub.
Explore before you publish gold
Run pandas notebooks on bronze and silver tables, visualize results, and save outputs back to the catalog.
Explore Lab notebooksTrain on gold datasets
Connect published gold tables to no-code ML, Forecast, and Loop — no export step required.
Explore Heimdall MLLake FAQ
- What is Heimdall Lake?
- Heimdall Lake is a cloud data catalog using medallion architecture: bronze for raw ingest (CSV, Excel, databases, unstructured zip archives), silver for curated views with filters and joins, and gold for model-ready datasets connected directly to Heimdall ML, Forecast, and Loop.
- How does Heimdall Lake compare to Databricks?
- Databricks is an enterprise lakehouse platform requiring Spark expertise and separate ML tooling. Heimdall Lake targets teams that want medallion data curation with built-in no-code ML deployment — simpler setup, no Spark cluster management, and native gold-to-model pipelines without exporting to another service.
- What data sources can Heimdall Lake ingest?
- Lake supports CSV and Excel uploads, database connections (PostgreSQL, MySQL, MariaDB, and more), and unstructured zip archives. Curated silver views and gold datasets feed directly into Heimdall's ML products.
- Does Heimdall Lake replace Amazon S3?
- Heimdall Lake is a managed data catalog with curation workflows, not raw object storage. Teams that need a full platform (ingest → curate → model → API) benefit from Lake's integrated pipeline; teams with existing S3 lakes still need separate ETL and ML tooling that Heimdall bundles natively.
Get your data into gold
Ingest into bronze, curate silver views, and publish gold datasets that connect to ML, Forecast, and Loop — no export hop.