This repo contains templates to build End-to-End Machine Learning Ops Using Below Stacks:

• Remote Storage engines (such as GCS, S3, Azure Storage, etc.)
• Data Versioning Tool: DVC by iterative.ai
• Mlflow A Machine Learning research lifecycle tool
• Seldon: An Opensource Kubernetes Framework to deploy machine learning models.

The Flow diagram below shows how we utilize Klops Ecosystems in our Data Science lifecycle.

• Clone this repo: git clone https://gitlab-engineering.koinworks.com/data-team/klopsec.git
• Change directory to this repo's root folder: cd klopsec
• Connect to your Kubernetes cluster. E.g (using GKE): gcloud container clusters get-credentials $CLUSTER_NAME --region $REGION --project $PROJECT_NAME
• run: ./install.sh
• The MLflow Tracking Server would serve under port 5000 and The Seldon Core Deployment under port 80

• run: ./install_monitoring.sh
• It will serve under port 9090.

All APIs and usages are already defined in the repository library here.

• Integrate MLflow Tracking Server within the cluster
• Adding support for Pod Monitoring
• Implement Authentication for MLflow Tracking Server
• Implement Authentication for Seldon Core
• Adding support for Https (TLS) connections
• Integrate with Model monitoring service
• Implement A/B Testing mechanism.

• Fork this repository.
• Do your changes / features.
• Ask for a merge request to the staging branch with a reviewer described in your merge request.

This repository were authored by Me

Apache License, Version 2.0.