Deploying Scalable ML models using K8s and TFserving/FlaskAPI