The original dbt model layer for a high-growth gaming product — and the ownership patterns that scaled the team from zero to eight engineers.
Problem
A high-growth daily fantasy sports operator had no analytics engineering function and no shared semantics in the warehouse. Analysts and data scientists were rebuilding the same joins on terabyte-scale transactional data.
Approach
Design the core dbt model layer for the gaming product, onboard the team to dbt Cloud, and split ownership by domain as headcount grew. Stay hands-on as architect while engineers owned vertical slices.
Architecture
Model layer
Raw transactional data through staging, marts, and semantic layers to every consumer.