Blueprints for Scaling: Practical Steps for Scaling Your Data Platform to Handle Exponential Growth
summary
Canva’s data platform grows to ~25 PB with ~400–450 contributors and ~7,000 PRs/year; platform teams build an internal provisioning tool called Tundra for Snowflake, adopt dbt, FiveTran and Bigeye, and implement CI checks (Buildkite-style) that prevent breaking changes, saving hundreds of thousands of dollars in remediation. The team decentralizes into domain-aligned teams vs a monolith, provides composable primitives and paved-road modules, and slashes PR environment spin-up from 45 minutes to under 5 minutes, driving faster time-to-insight.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:08Scaling pains in data platforms
- 2:08Canva's platform growth
- 4:16Aligning platforms to business value
- 8:10Choosing platform topology
- 12:21When to use data mesh
- 15:28Composable components for platforms
- 18:59Driving adoption with incentives
- 23:53Measuring platform performance
- 27:49Recap and scaling analogy
related talks
35mBuilding a data analytics platform on strong opinions held loosely
23mFrom Start-Up to Scale-Up: Building a Data Team at a High-Growth Global Ed-Tech Company
38mKeeping pace with data infrastructure so it doesn't outrun you
19mKickstarting a greenfield data project
17mStrengthening Security and Compliance in Our Data Platform
23mData Team as an Optimisation Problem
40mThe Tools Empowering 5000 Employees
29mData strategies to enable ML and analytics at scale