Use Kubeflow Pipelines to Deploy Your On-Prem ML Workloads to Google Cloud

If you are wondering how to bring your on-prem ML workloads to a scalable, portable, composable and secure production platform, the open source Kubeflow pipelines is your answer. We demonstrate an easy step by step process with ML models from scikit-learn, xgboost and tensorflow ML frameworks. We will show how to create an end to end ML pipeline on the Google Cloud including data prep, hyperparameter tuning, model training, model deployment, prediction, explanation and training orchestration. The solution can be extended to the Anthos framework for a full multi-cloud deployment.

Chanchal Chatterjee, Ph.D, held several leadership roles in machine learning, deep learning and real-time analytics. Chanchal is currently leading Deep Learning software at NVIDIA server platforms. In the past, he led many ML/AI projects at Google Cloud Platforms with Fortune 500 customers in several verticals. Chanchal received an Outstanding paper award from IEEE Neural Network Council for adaptive learning algorithms recommended by MIT professor Marvin Minsky. Chanchal also has a book titled Adaptive Machine Learning Algorithms with Python which has exceeded 15K downloads.