SRE and ML: Why It Matters
summary
Machine learning systems are becoming part of core infrastructure: adoption rises, models run critical paths (recommenders, fraud, autoscaling), and operations teams must manage training, deployment and fragile, data-sensitive pipelines. ML reliability proves hard: models need labeled data, drift breaks behavior, and heuristics often beat ML except in cases like autoscaling worked demos. Expect new roles—model quality SRE and data reliability engineers—and operational focus on measurable model quality, instrumentation, and governance.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 2:37Lambda
- 4:59Dave
- 10:57TMU
- 13:29ML does matter
- 14:37Pause and breathe
- 15:50Managing ML in production
- 20:38How hard is ML
- 27:25Hype and Reality
- 27:37Gartner Hype Cycle
- 30:42ML vs AI
- 32:12ML Ops
- 37:33Model Quality
- 39:30ML is data sensitive
- 40:42I can ML
- 41:21The future
- 41:36Future predictions
- 43:16Future reading
related talks
44mSRE for ML: The First 10 Years and the Next 10
39mAn SRE Approach to Monitoring ML in Production
26mCan SREs become AI/ML engineers?
18mReliable Maintenance of Machine Learning Models
46mI know Machine Learning but... am I ready for the real world?
32mReliable Machine Learning
06mMachine learning is about your data and deployment
25mAutomating Operations with ML