Case study
Building a data platform from scratch
Designing a cloud-native data platform end to end, from the target architecture to the warehouse, and the team to run it.
Context
The organization needed a modern data platform that every data team could build on.
Challenge
There was no shared foundation. The platform had to be reliable, easy to reason about, and run by a small team that was still being formed.
Approach
- Defined the target architecture, with all infrastructure written as code.
- Delivered the core layers: data orchestration, a data lake and lakehouse, the data warehouse, and modeling in dbt.
- Built and led the data engineering team that runs it.
Outcome
The platform went live with 15+ production-grade ingestion pipelines, designed and shipped at speed.
What I learned
The hardest part of a data platform isn’t the architecture. It’s earning trust, one reliable number at a time.