Start the head node with
ray start --head
Build a docker image with all necessary deps for running the application. Start workers with
ray start --address ray-head-ip:port
Build a config file for the serve application with:
serve build app_file_name:application -o deploy_config.yaml
Deploy the application with
serve deploy deploy_config.yaml