#376 Rethinking the Data Stack in the age of AI with Tristan Handy, President of Fivetran + dbt Labs
Summary
Tristan Handy, President of the newly merged Fivetran and dbt, discusses the evolution of the data stack toward open standards like Iceberg to reduce vendor lock-in. He explains how AI agents are transforming data teams by shifting roles from report creation to platform maintenance and semantic layer governance. The episode highlights the critical need for high-quality, governed data definitions to ensure accurate conversational analytics. Handy also shares his approach to using agents for complex research tasks, emphasizing structured context engineering and traceability to minimize halluc…
Topics discussed
DataCamp AI Skills Sponsorship
The AI Data Context Problem and Guest Intro
Rationale Behind the dbt and Fivetran Merger
Explaining Fivetran Ingestion vs dbt Transformation
Competing with Snowflake and Databricks
Open Data Infrastructure and Avoiding Lock-in
Designing a Flexible, Multi-Store Data Stack
Managing Complexity and Vendor Choice
AI Coding vs Data-Driven AI Use Cases
The Rise of Conversational Analytics
The Critical Role of the Semantic Layer
Apache OSIE and Standardizing Semantics
Shifting Data Team Roles: Platform vs Enablement
The Future of Data Job Titles and Skills
Career Progression and the Junior Talent Gap
How Tristan Adapts to AI and Agent Tools
Using Agents for Research and Literature Reviews
Tips for Effective Context Engineering
Recommended AI Resources and Closing
Listen ad-free on Castria