Cloud Computing for ML covers the concepts and how the three providers map onto each other. Deploying Models on AWS & Azure covers real model-serving deployment methods. This page is the command reference for the three CLIs side by side, for the tasks you need in all of them: auth, compute, storage, and managed Kubernetes.
Authentication
# AWSaws configure # interactive: access key, secret, region, output formataws sts get-caller-identity # confirm which identity is currently activeexport AWS_PROFILE=my-profile # switch profiles without re-running configure# Azureaz login # opens a browser for interactive authaz account set --subscription <name-or-id> # pick which subscription commands targetaz account show # confirm the active subscription# GCPgcloud auth login # interactive user authgcloud auth application-default login # credentials for SDKs/client libraries, separate from the CLI's own authgcloud config set project <project-id> # pick which project commands target
# AWSaws iam create-role --role-name my-role --assume-role-policy-document file://trust-policy.jsonaws iam attach-role-policy --role-name my-role --policy-arn arn:aws:iam::aws:policy/ReadOnlyAccessaws iam list-roles# Azureaz role assignment create --assignee <principal-id> --role "Contributor" --scope <resource-scope>az ad sp create-for-rbac --name my-service-principal # create a service principal for CI/automation# GCPgcloud iam service-accounts create my-service-accountgcloud projects add-iam-policy-binding <project-id> --member="serviceAccount:my-sa@<project-id>.iam.gserviceaccount.com" --role="roles/storage.objectViewer"