Description
Over the past few years, we have seen leaps and strides in ML and most recently generative AI. Companies and software teams are rushing to enhance, rebuild, and create new software offerings with this new intelligence. As they innovate and create delightful new experiences for their customers new challenges arise. Understanding how these applications work and how to use state-of-the-art infrastructure tools like Kubernetes will help organizations and professionals succeed with this new technology.The book covers essential technical implementations from ML fundamentals through advanced deployment strategies, focusing on practical patterns. Core topics include Kubernetes-native GPU scheduling and resource management, MLOps pipeline architectures using Kubeflow/MLflow, and advanced model serving patterns. It details data management architectures, vector databases, and RAG systems, alongside monitoring solutions with Prometheus/Grafana. Finally, we will look at some advanced concerns for production in the realm of security and data reliability.After reading this book, you will be equipped with a broad knowledge of the end-to-end generative AI pipeline and how Kubernetes can be leveraged to run your generative AI workloads at scale in the real-world.
Key Features
? Learn how Kubernetes can help you run your generative AI workloads.
? Using hands-on examples, you will work with real-world foundational models and a variety of tools and capabilities in the K8s ecosystem.
? A broad survey of both generative AI and Kubernetes in one book.
What you will learn
? How to evaluate and compare models for new applications and use cases.
? How Kubernetes can add reliability and scale to your AI applications.
? What does an AI delivery pipeline contain and how to start one.
? How AI models encode words and work with natural language.
? How prompting and refinement techniques can improve results.
? How to use your own data to augment AI responses.
Who this book is for
This book is for teams building new applications or new functionality with generative AI, but want to better understand the infrastructure needed to bring their AI applications to production. This book is also for shared services, infrastructure, or cybersecurity teams who provide platforms and infrastructure for application, or product development.
Table of Contents
1. Introduction to Generative Artificial Intelligence
2. Kubernetes for Generative AI
3. Introduction to Foundational Models on Kubernetes
4. Working with Foundational Models
5. Process and Pipelines
6. Process and Pipelines on Kubernetes
7. Managing Data for Generative AI
8. Refining and Improving Results
9. Observability and Monitoring
10. Securing ML/GenAI Pipelines on K8s