Architecting and deploying machine learning for private cloud

Description

Laura Schornack and Derek Ferguson walk you through the conception, design, and implementation of an end-to-end machine learning system for financial pattern recognition in the private cloud. Starting with just a basic theoretical knowledge of machine learning and containerization, you’ll walk away with a complete understanding of both the challenges in private cloud machine learning and the best solutions.

Laura and Derek begin by discussing the underlying machine learning concepts to be used in the solution. They describe the case—recognizing latent patterns in financial data—and introduce the technologies in the solution (TensorFlow, Kubernetes, and Kubeflow). Laura and Derek then walk you through the overall architecture of the solution, including primary stores and conduits of data into the system, backbone processing for training of the ML models, and execution capacity for predictions against these models. They conclude with a deep dive into the unique challenges around establishing a full ML solution in a private cloud.

Topics include:

• Bringing tools primarily designed for public cloud in-house
• Achieving optimal neural network performance in GPU-restricted environments
• Establishing secure persistent volumes on Kubernetes for confidential data
• Debugging complex ML models across multiple containers
• Open source contributions in Fortune 50 enterprises