MLOps Demo

How to set up

Locally

Install Python 3.11
Execute pip install -r requirements.txt
Install Tilt
Make sure you have a local Kubernetes Cluster with either docker-compose, rancher-desktop or minikube
Execute tilt up

Google Cloud (Currently not fully supported use tag 1.0.0 for full support)

First of all, you need the Google Cloud Credentials JSON for the Computing Engine Default Service Account. If you don't have one, create it here. The key needs to be named key.json and put into the backend Directory.

Create a Vertex AI Endpoint

Go to Vertex AI Model Registry
Select newest Version
Go to 'Deploy & Test'
Click on 'Deploy to Endpoint'
Enter a Name for the new Endpoint
Go to 'Model Settings' and select 'n1-standard-2' for Machine Type
Click on 'Deploy'
Go to Vertex AI Endpoints and copy the ID of our newly created Endpoint.

Create Kubeflow Pipelines Instance

Go to Kubeflow Pipelines
Select 'Configure'
Create a new cluster by selecting europe-west1-c as zone and ticking the box at 'Allow access to the following Cloud APIs'
When the cluster is created, click on 'Deploy'
After Deployment, go to AI Platform Pipelines
Click 'Open Pipelines Dashboard' on you newly created Instance
Copy the URL

Prepare Dockerfile

Go to backend/Dockerfile
Set ENDPOINT_ID to our Vertex AI Endpoint ID
Set KUBEFLOW_URL to our Kubeflow Pipelines Instance URL
Start the Docker Container

After the preparation, you can run the Webapp by running npm start in the frontend directory