uv(brew install uv)- AWS CLI (
brew install awscli)- Create an AWS profile called
mlops-club
- Create an AWS profile called
nodejsandcdkCLI (brew install node; npm install -g aws-cdk)- Docker Desktop (
brew install docker)
And, if you like AI tools:
-
Amazon Q CLI. (I like it) -
brew install --cask amazon-q^^^ I've found
q chatto be good at AWS-based tasks.
./run# activate an mlops-club profile with broad levels of permissions to deploy cloudformation stacks
aws sso login --profile mlops-club # for example
# bootstrap your aws account (only necessary once ever per account)
./run cdk-bootstrap
# modify the infra.py file to use a hard-coded bedrock inference profile for the model in your aws account
# deploy the application
./run cdk-deploy
# clean up the application
./run cdk-destroy