Implement MLOps practices with Amazon SageMaker
summary
Shows how SageMaker Pipelines automates end-to-end ML CI/CD with project templates, cross-account artifact access, and a visual pipeline UI; builds a demo pipeline (project creation ~2 minutes, full pipeline ~10 minutes) that runs preprocessing, training, conditional evaluation and deployment into staging via a model registry with manual approvals; example metrics show an RMSE of RMSE 2.3 on validation; code commit triggers retraining (train/test split changed 70→75%) and the published flow works as shown.
AI-written summary · about
chapters (12)
- 0:00Intro
- 2:10What is SageMaker
- 5:09How are models trained and built
- 10:10SageMaker Pipelines
- 13:37ML Workflows
- 15:25Troubleshooting
- 16:47Managing Models
- 18:32MLOps Templates
- 20:00Lineage
- 21:01Demo
- 28:55Commit to pipeline
- 32:09Resources
related talks
34mImplement MLOps practices with Amazon SageMaker
29mImplementing MLOps practices with Amazon SageMaker
29mHow to create fully automated ML workflows with Amazon SageMaker Pipelines
45mAutomate your ML workflows with pipelines
36mImplementing MLOps practices on AWS
43mMachine Learning Ops: DevOps for Data Science
33mBuild, train, and deploy ML models using Amazon SageMaker
33mBuilding & Managing MLOps Pipelines in the Cloud