Becoming a Data Head, a reflection on the trade offs and decision
summary
Claims concrete tradeoffs: builds mission-driven hiring to recruit people, runs a platform data team not point-to-point, and accepts that ~50% project failure is common. Names tools: Jupiter, conda, pandas, numpy, scikit-learn, mlflow, Docker, SageMaker, and bouquet dashboards used to ship results. Gives numbers: led data teams to about 50, then 10, now ~20 data scientists inside a 300-person company. Runs demo-day Fridays to surface wins and prefers reproducible, open workflows that actually get used in production.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 2:25Founder dilemmas and trade-offs
- 5:36Defining mission for data teams
- 9:11Innovation vs strategy
- 17:59Strategic data team building
- 28:40Economic principles in data projects
- 43:29Roles and resource management
- 44:13Q&A
related talks
30mDevOps For Data Teams
42mBuilding a Data Driven Organization
32mWhy Most Data Projects Fail and How to Avoid It
53mThe State of Data Engineering
26mThe Software Engineering Part of Data Science
19mData science demystified
01h 00mThe challenges and frontiers of data science in Europe
21mPragmatic data organization leadership