Which we did said no. that was right answer as deployment would have needed money as even on ec2 of aws, big ML model can't be run. more ram will be needed. Sagemaker is ultimately used but that too is paid