Data engineeringAdvanced complexity
Warehouse Foundations
Builds the modelled customer data layer in your warehouse (dbt-style): clean, tested, documented tables at customer grain that every model and sync reads from.
A trustworthy customer data layer, so the models are built on modelled data, not raw exports.
Who it's for
Clients whose warehouse holds raw source data but no clean, customer-grain model. The prerequisite the models quietly assume.
The engagement
What you get
- A modelled customer data layer (dbt or equivalent) at customer grain
- Tested, documented transformation pipelines from source to model
- A data dictionary and lineage for the modelled tables
What Fuse does
- Map the source systems and agree the target customer data model
- Build and test the transformation pipelines
- Document the model, lineage and refresh schedule
What we need from you
Data & access
- A cloud data warehouse
- Access to source data
Your responsibilities
- Provide access to source systems and the warehouse
- Confirm business definitions with data and finance owners
Outcomes and proof
- Every model and sync reads from one clean, tested source
- Faster model builds, fewer data surprises
- A documented, governed customer data layer
Assumptions
- Source data is available in, or landable to, the warehouse
Out of scope
- Source-system extraction tooling or licences
- The models themselves, which are separate services