Automating Production Production Level Machine Learning on AWS
summary
Framework named "model factory" claims to cut ML deployment from months to weeks by reducing 25+ manual steps to ~10 and automating promotion across dev/QA/prod. It packages models with a model DSL manifest, tracks experiments with MLflow + SageMaker + CodePipeline/CodeBuild, and uses Conda/Docker, ECR, EMR, EKS/Argo or SageMaker for serving. A water-utility pilot integrates existing models, achieves reproducible experiment IDs, faster training-to-production turnaround, and consistent monitoring via CloudWatch/Datadog and data checks.
AI-written summary · about
chapters (9)
- 0:00Introduction to ML automation
- 3:52Why ML operations matter
- 8:25AWS ML stack overview
- 11:55Model Factory framework
- 15:23Development lifecycle comparison
- 24:06Model deployment process
- 26:54Governance and security
- 29:00Customer success story
- 31:49Q&A and closing
related talks
25mDesigning an AWS Well Architected Framework for AI/ML
24mFrom Model to Production: Automating the ML Lifecycle with CI/CD
01h 05mMachine Learning Models and Productionization
44mPragmatic Machine Learning in the Cloud
43mMachine Learning Ops: DevOps for Data Science
32mProductionize Machine Learning on Azure
26mManaging ML Models @ Scale
43mDesigning and Building Complex Machine Learning Projects