Panel Discussion
summary
Debate over when to add guardrails gives concrete rules: treat data contracts as codified APIs (schema, SLAs, ownership) when data is a product; delay heavy modeling until product–market fit but store schemas and ownership to avoid painful rewrites; use GitHub for iterative governance and quantify downstream outages in dollar impact to justify fixes; cloud-hosted DuckDB via Motherduck is positioned for lightweight analytics; aim for minimal compliance (SOC 2) templates and small Wikis first.
AI-written summary · about
chapters (13)
- 0:00Welcome
- 3:55Data engineering for startups
- 4:59Data contracts in startups
- 7:04Data modeling for startups
- 8:28Growing data platform
- 13:21Q&A on governance
- 18:02Governance versus speed
- 21:28Why stay in data engineering
- 26:48Career advice in tech
- 28:54People problems in data engineering
- 31:17Cultural change and data contracts
- 33:03Survival tips for solo engineers
- 34:06Closing
related talks
20mData Contracts: Keep Your Weekend Work-Free!
29mHow to treat your data as a product
39mPanel Discussion
41mPanel Discussion
54mBehind the Hype - Architecture Trends in Data
27mLive in the Data Wild West: The Data Contracts Sheriff
47mKeynote - The Evolving Role of Data in Software Development
33mPanel Discussion