Three sources feed six custom point-to-point Python jobs that land directly into a finance dashboard

A medium Pipeline Design interview practice problem on DataDriven. Write and execute real pipeline design code with instant grading.

Domain

Pipeline Design

Difficulty

medium

Problem

Three sources feed six custom point-to-point Python jobs that land directly into a finance dashboard, a product dashboard, and an ML feature store. There is no shared raw layer, no shared curated layer, custom-coded ingestion, one orchestrator owning everything, and no undercurrents. Apply the five redesign steps the section walks through to turn this into a layered architecture with named contracts, bought commodity tooling, and split DAGs at team boundaries. Postgres ingestion stays custom; only commodity SaaS connectors are bought.

Practice This Problem

Solve this Pipeline Design problem with real code execution. DataDriven runs your solution and grades it automatically.