How a Health Media Co Built 140 Data Models in 6 Weeks
A healthcare media CEO handed off her Snowflake and dbt buildout to Claude Code — 99% AI-written, 140 production data models shipped in six weeks instead of nine months.
99%
Of code written by Claude Code
The Challenge
A healthcare media and events company came to Wallabi asking for an n8n email automation expert. After the first build, the CEO said it wasn't what she needed. After a second pivot to data access, she said that wasn't it either. What surfaced on the third attempt: the company had no coherent data architecture. Data was dirty, scattered across systems, duplicative — and before any automation could work, they couldn't even answer a basic question about who was in their database.
What They Built
Wallabi built a full data infrastructure using Snowflake, dbt, Fivetran, and their Wallabi platform — with Claude Code writing 99% of the code and documentation throughout. A CI/CD process built around Claude Code let them roll out new data models rapidly, stacking context and learnings directly into the repository so every team member benefited from every previous session. In six weeks, they built 140 data models on a production-ready medallion architecture — work that would have taken 6–9 months with a traditional team. On top of it: Tableau for reporting, Snowflake Intelligence for natural language queries, and reverse ETL pushing clean data back to their CRMs. Nobody expected the final outcome: Wallabi is now directly shaping the company's board narrative and 2026 growth strategy — capabilities the CEO didn't know she needed when she first knocked on the door.
Wallabi's engagement began as an n8n email automation request. After two failed pivots, the real problem emerged: the client had no coherent data architecture. Data was dirty, duplicated, and scattered across disconnected systems. Wallabi scrapped the original scope and redesigned the engagement entirely. They built a modern data stack from scratch — Snowflake as the warehouse, dbt for transformation, Fivetran for ingestion, and the Wallabi platform as the application layer. A CI/CD process was built around Claude Code, which wrote approximately 99% of the code and documentation throughout. Prior session context was stacked directly into the repository so every team member benefited from accumulated learnings. This enabled Wallabi to roll out nearly 140 independent data models in six weeks — work that would have taken six to nine months with a traditional team. On top of the warehouse, they added Tableau for reporting and Snowflake Intelligence for natural language queries. Reverse ETL pipelines pushed clean data back to the client's CRMs. The final outcome: Wallabi's work became a direct input to the client's board narrative and 2026 growth strategy.
Infrastructure
Snowflake (cloud data warehouse) • dbt (data transformation and modeling) • Fivetran (data ingestion/ELT) • Wallabi platform (application layer) • Tableau (reporting and dashboards) • Snowflake Intelligence (natural language query layer) • Claude Code (AI-assisted development — ~99% of code)
Integration Points
Source systems → Fivetran → Snowflake (raw/bronze layer) • Snowflake → dbt transformation → medallion architecture (silver/gold layers) • Gold layer → Tableau dashboards for reporting • Gold layer → Snowflake Intelligence for NL querying • Gold layer → reverse ETL → client CRM and marketing automation tools
Impact
Claude Code wrote all code and documentation for the data architecture, and 100% of the code for the intelligent applications on top
Built in 6 weeks on a production-ready medallion architecture — work that would have taken 6 to 9 months with a traditional team
Business teams that previously waited weeks for a single data report now get answers in real time
Implementation Complexity
Best Fit For
Mid-market companies in services, media, or events that know they need to do something with AI but keep hitting dead ends — especially those who have already bought the right tools and still aren't seeing results from them.
