Business-owned metadata

The Culina framework.

One batch ETL engine. Business-owned rules. A clear boundary between the definition of the work and its execution.

One operating model, three moves.

Keep the story simple: define the work, run it consistently, and keep evidence visible.

Define

Capture business meaning, source rules, and pipeline contracts.

Operate

Run the same structure through orchestration, checks, and logs.

Prove

Keep validation and handover evidence tied to every delivery.

A standardized batch ETL engine

The Culina framework separates the definition of a data job from its execution. Approved metadata describes sources, dependencies, transformations, and runtime policy. Orchestration creates batch executions and checks which jobs are ready before dispatching work to the shared runtime. This is a standardized batch ETL operating model, not a bespoke platform build or a streaming engine.

Each execution has its own queue state and recorded outcome, separate from the job definition.

Business meaning stays in the contract

Requirements and discovery establish what the data must represent. Modeling defines record grain, keys, columns, and persistence behavior before those decisions become metadata. Business-owned rules therefore remain explicit rather than being hidden inside workload code.

The shared runtime supplies reusable execution behavior. Environment configuration supplies platform-specific inputs. Keeping those responsibilities separate makes the approved definition reviewable without confusing a business decision with a storage location or a runtime implementation.