Welcome + Opening Remarks: The State of Production MLOps in the Cloud Native...
summary
Argues that the ML lifecycle begins at deployment and lists concrete controls: an architectural blueprint separating training, artifact store and inference; standardized ML metrics (accuracy, precision, recall) plus drift detection and explainability deployed as infrastructure; metadata linking inference back to training for relabeling; security controls across every lifecycle stage; organizational patterns shifting from project to product with a suggested ratio of one MLOps engineer per multiple pipelines to scale hundreds of models.
AI-written summary · about
chapters (18)
- 0:00Intro
- 2:51Model ende begins once trained
- 8:02Consolidating Practical AI Ethics
- 9:28Consolidating Accountability Struct.
- 13:01AI Open Source Libraries as Policy
- 14:47Architectural Blueprint Convergence
- 18:11Converging Into Cannonical Stack(s)
- 19:07Maturing Monitoring Areas in ML
- 20:51Evolving to Observability By Design
- 21:48From Model-Centric to Data-Centric
- 23:30Robust DataOps/DataMesh Evolution
- 24:24End to End Metadata Interoperability
- 26:39Demand for Secure End to End MLOps
- 29:45Mindset From Projects to Products
- 31:24Map Tech Outputs to Biz Outcomes
- 32:36Cross Func. Capabilities By Design
- 33:27Exploring Organisational Ratios
- 35:46Iterative Organisational Growth
related talks
43mNavigating the MLOps Journey: Key Considerations for Successful Implementations
41mTHE STATE OF MLOPS - MACHINE LEARNING IN PRODUCTION AT ENTERPRISE SCALE
47mMachine Learning in Production: An Intro to MLOps.
13mMLOps: From AI Research to Production-Ready Systems
47mMLOps tools to automate ML deployment
01mMLOps: Applying DevOps to ML
45mThe Culture, Architecture and Technology of MLOps
20mBringing CI/CD Practices to Machine Learning with MLOps