Simplify, speed and improve development with DevOps
summary
Argues that DevOps practices can handle AI/ML end-to-end and demonstrates a working pipeline: code and Jupyter notebooks in git, Azure Repos/GitHub for versioning, Azure DevOps boards and pipelines to build and release, Azure Machine Learning for training and model management, a container registry and deployment to Azure Container Instances. Gives surprising scale metrics—80,000 deployments per day, 4.5 million monthly builds and 5 million daily work items—and shows a web app calling a containerized handwritten-digit model that updates successfully via CI/CD.
AI-written summary · about
chapters (10)
- 0:00Start
- 1:59What is DevOps?
- 6:15Elements of DevOps
- 8:42Developers are familiar with agile processes for some years
- 10:26Example of a DevOps organization Microsoft
- 12:20DevOps pain points from Data Scientists
- 16:57How to choose a platform for ML/DevOps projects?
- 22:10Azure Machine Learning service
- 22:38Model management and deployment
- 23:24Azure DevOps
related talks
57mDevOps for Machine Learning
52mDevOps for AI in the Microsoft World
56mDevOps for Machine Learning
01h 31mAccelerate Business Delivery With AI and DevOps
42mCI/CD for Machine Learning
24mFrom Model to Production: Automating the ML Lifecycle with CI/CD
52mHands-on Azure DevOps (...)
06mIntroduction to Azure DevOps