Story
Databricks
SandvikSandvik Brings Feature Pipelines onto Databricks
Sandvik, an industrials organization in Sweden, uses Mosaic AI from Databricks to support applied machine learning for data scientists and ML engineers.
Data scientists and ML engineers at Sandvik review feature pipelines in Mosaic AI rather than in personal files.
“Inside Sandvik, applied machine learning used to depend on whoever still had the latest file.”
“Data scientists and ML engineers needed a system that would still make sense after the original project team moved on.”
Independent write-up. Figures come from public sources or are illustrative.
Value results
| Category | Value result |
|---|---|
| Capability | Feature pipelines can be reviewed without waiting on a personal export |
| Capability | Named workflow replaces ad hoc routing for feature pipelines |
| Capability | Applied machine learning stays visible to adjacent teams through Mosaic AI |
Story
Inside Sandvik, applied machine learning used to depend on whoever still had the latest file. That pattern is common in industrials groups working out of Sweden. Data scientists and ML engineers needed a system that would still make sense after the original project team moved on.
Sandvik uses Mosaic AI from Databricks as the working layer for feature pipelines. Databricks provides a lakehouse platform for data engineering, analytics, and AI, with Unity Catalog and Mosaic AI for governed models. The practical change is simple: applied machine learning has a home, and reviews happen there instead of in a forwarded thread.
Nothing in this writeup invents a savings number. What Sandvik gets from Databricks is a durable place to run applied machine learning and a way for data scientists and ML engineers to see the same feature pipelines at the same time.
