How to create a high performance data team: lessons learned from the field
summary
Cloud-first teams choose small multidisciplinary groups (around seven to nine, think two pizzas) and aim for engineering maturity: automated CI/CD on Azure, a Synapse data platform, and explicit non-functional requirements. A legacy Robo-advice product fails when independent test and deployment teams force quarterly releases instead of weekly. A Wealth Management client gets a new Synapse platform in three months, adds a domain expert and grad, later hires a software engineer, keeps a retained brain trust two hours per week, and delivers multiple data products successfully.
AI-written summary · about
chapters (9)
- 0:00Start
- 3:04Conway's Law & homomorphic force
- 8:04small multi-disciplinary team
- 9:52Attitude over aptitude
- 10:45First hires?
- 14:23product mindset
- 15:15Space created to focus on sustaining value
- 15:40engineering maturity
- 16:43organisational readiness
related talks
22mBuild Smarter, Don't Just Build
50mMove Fast, Think Big
38mBuilding Effective Data Teams: Strategies for Success
30mBuilding flexible data teams to improve product delivery
29mBuilding successful data teams in the realities of 2023
30mMoving Fast At Scale
28mMaking do: Scaling expertise on teams you can't grow
44mStories from the trenches: Building product teams ...