This project has the sole purpose of serving as a template for a kubernetes deployment.
To serve the model I used the pickle stored in the models folder in this repo.
The application was deployed to a Container Registry, and the image (as can be seen in Dockerfile) was loaded in the kubernetes pods exposing the API (as can be seen in deploy.py)
For the purposes of deploying in Kubernetes the code for this project was not made available in scripts
I used the Dockerfile to deploy the application via the deploy_code.py script and ran the deploy.py to expose the API in order to serve the model.
This project was tested in an Azure environment and for that, the folder azure-pipelines was used for Continuos Integration and Continuos Deployment
In the kubernetes folder lies all the configuration of the kubernetes cluster used