Foundations for ML at Scale: Where to Start in Your Organization
summary
Start projects that directly move business metrics and prioritize by projected ROI, treating ~10–30% of work as venture bets for high-risk innovation. Invest heavily in data pipelines and labeling (the advice: spend ~6 of every 10 dollars on data), instrument experiments end-to-end with tools like weights and biases, aim for median-time-to-deploy reductions, get executive sponsorship, and expect small teams (two engineers) to ship features like People You May Know that drive millions of users.
AI-written summary · about
chapters (12)
- 0:00Introduction and speaker intro
- 1:05Why ML platform success matters
- 3:10Align ML with business goals
- 6:18Executive sponsorship and data access
- 7:21Prioritizing high-impact projects
- 11:57Investing in data infrastructure
- 13:20Data quality and labeling strategies
- 15:50ML engineer productivity
- 17:55Machine learning testing
- 20:00MLOps automation essentials
- 21:27Monitoring model performance
- 22:50Communication and internal training
related talks
27mKeys to Building Machine Learning Systems
28mPractical steps to establish and scale your data & ML team
32mWhy ML Projects Fail and How to Ensure Their Success
42mMasterclass: The right path to ML
45mStart your AI/ML Projects Right
32mAn Only One Step Ahead Guide for Machine Learning Projects
30mBuilding Teams and Culture that Support ML Innovation
32mMaking Models Work: How an optimized ML lifecycle can drive better ROI