Building From Zero: Architecture, Trust, and the Semantic Layer
In this episode, Kevin Sampson, the first data hire at Vertex Service Partners, joins us to talk about what it takes to build a data and analytics platform from scratch, and why AI without a semantic layer is a tempting shortcut that breaks inside almost every real business.
We cover:
- Why building from zero is a gift, not a burden, inside a traditional industry
- How dashboard sprawl is a window into the real state of your data architecture
- Why most data teams are conflating insights with answers, and what AI is making worse
- Why AI without a semantic layer only works in a trivially simple business, and why almost no business is trivially simple
- How a single data engineer scales by investing in foundations before reaching for AI
This is a grounded conversation for data engineers and analytics engineers thinking about the foundations that hold up when AI lands on top of them.
The hot take
This episode also debuts our new format. Every guest faces a hot take. One question they have never heard before, designed to challenge how they think on the spot. For Kevin, the hot take was about whether AI without a semantic layer can really work. His answer is the kind of thing every data engineer should hear before reaching for an LLM.
About Kevin Sampson
Kevin is the first data hire at Vertex Service Partners, a private equity backed roll up of residential and commercial roofing brands. Before Vertex, Kevin spent four and a half years at Amazon as a business intelligence engineer inside their logistics organization.
About the host
Aaron Phethean hosts the Data Matas Podcast, a series of conversations with practitioners building modern data stacks.
