Scaling data science in the enterprise
summary
RBC standardizes on Dataiku as a single hybrid platform so 200+ data scientists share code, models and pipelines; a concrete project ingests PDFs and spreadsheets to automate audit controls automation, packages outputs as REST APIs and dashboards, and moves seamlessly from central build to embedded team ownership. The hub‑and‑spoke pattern yields a reusable internal data & AI product, reduces time‑to‑value, supports hybrid on‑prem/cloud constraints, and is already in production delivering measurable ROI.
AI-written summary · about
chapters (14)
- 0:00Introduction to scaling
- 1:47Why data science is hard to scale
- 3:53RBC’s data science structure
- 6:43Models vs analytics outputs
- 8:48Human vs machine consumption
- 9:30On-prem vs cloud at RBC
- 11:36Audit automation success
- 18:31Post-COVID trends in AI
- 23:03Commoditization of skills
- 25:29Gaining executive buy-in
- 26:29Next steps for RBC’s AI
- 27:52Measuring business value
- 29:38Hub and spoke scaling tips
- 31:02Closing and thanks
related talks
49mBuilding a Data-Driven Bank
15mHow AI and ML unlock insights for financial services teams
34mThe Data Analytics Platform or How to Make Data Science in a Box Possible
22mScaling analytics and AI: It's mostly about culture
29mBecoming a data-driven enterprise
52mElevate your data and AI governance with Databricks Data Intelligence Platform
22mBuild Smarter, Don't Just Build
27mDemocratising data-driven decisions with self-service tools