Welcome to AI500!
AI500 is a MLOps enablement, taking you through the inner loop all the way out to a fully automated system that trains and deploys models based on triggers such as data versions, git changes, and metrics.
Here you can find the full enablement along with everything needed to set up an environment and run through it.
It currently consists of these repositories:
- lab-instructions - The AI500 instructions, rendered here.
- deploy-lab - All scripts needed to set up the baseline environment.
- jukebox - The data/ML persona notebooks and pipelines.
- jukebox-ui - Contains the UI used to send requests to the model.
- mlops-helmcharts - The charts needed for the student to deploy outer loop components.
- mlops-gitops - The app of apps, used to deploy and manage the components in mlops-helmcharts.
- happy-safe-model - A not very safe model for demonstrating how to scan models.
