MLOps from the edge to on demand visualization
summary
Argues that building MLOps on OpenShift + AWS requires concrete choices: uses Ansible for infrastructure-as-code, builds auto scaling hooks to safely scale OpenShift nodes, and creates a Cloud Control API to iterate templates; pipelines ingest edge data into S3, catalog with AWS Glue, query with Athena, train and serve via SageMaker or Open Data Hub containers. Describes the pilot phase as messy, then reproducible once decision records, CI/CD and workstation-on-demand hardware are in place.
AI-written summary · about
chapters (6)
- 0:00MLOps introduction
- 3:54Designing MLOps success
- 10:00Edge to visualization
- 17:49MLOps platforms
- 24:52Constructing a machine learning pipeline
- 28:51Summary
related talks
33mBuilding & Managing MLOps Pipelines in the Cloud
36mImplementing MLOps practices on AWS
47mIntroduction to MLOps at the Edge
53mExperience with an end-to-end MLOps pipeline: Practical insights and 3D architectural perspectives
13mMLOps at Scale
36mDataOps and MLOps: NLP Development and Deployment in Kubernetes made easy
39mContinuous Delivery for Machine Learning — The Future of MLOps
41mTHE STATE OF MLOPS - MACHINE LEARNING IN PRODUCTION AT ENTERPRISE SCALE