دانلود کتاب آموزش عملی MLOps در Azure – خودکارسازی، ایمن‌سازی و مقیاس‌بندی گردش‌های کاری ML با Azure ML CLI، GitHub و LLMOps

Machine Learning Operations (MLOps) is an emerging discipline that brings together machine learning, DevOps, and data engineering to streamline and automate the end-to-end lifecycle of machine learning models—from development and experimentation to deployment and monitoring. This book introduces MLOps in a practical, scenario-driven way, with real-world examples using Azure ML, GitHub Actions, and cloud-native services. It aims to help you operationalize machine learning models efficiently and reliably in enterprise environments. The book concludes by exploring the latest trends in LLMOps—applying MLOps to large language models such as GPTs. This book is written for DevOps engineers, cloud engineers, SREs, and technical leads who are involved in deploying and managing machine learning systems. It also serves project managers and decision-makers looking to understand MLOps processes and best practices. You are expected to have a working knowledge of the following: • Machine learning concepts (model training, evaluation, data preparation) • Cloud computing (Azure, AWS, or GCP) • Software development tools such as version control, testing, and CI/CD • Python programming A background in DevOps is especially helpful, as this book builds on DevOps principles and extends them to ML workflows. Chapter 1, Understanding DevOps to MLOps, introduces DevOps fundamentals and transitions into MLOps practices such as faster experimentation, deployment, and model governance across cloud platforms. Chapter 2, Training and Experimentation, guides you through creating ML workspaces, tracking experiments, and optimizing models using hyperparameter tuning. Chapter 3, Reproducible and Reusable ML, focuses on building repeatable ML pipelines and managing environments to ensure consistent and efficient ML development. Chapter 4, Model Management (Registration and Packaging), covers model registration, packaging, versioning, and deployment strategies to support the full model lifecycle. Chapter 5, Model Deployment: Batch Scoring and Real-Time Web Services, explores how to implement scoring jobs for batch processing and real-time prediction using scalable cloud services. Chapter 6, Capturing and Securing Governance Data for MLOps, delves into governance, lineage tracking, compliance, and security of ML workflows. Chapter 7, Monitoring the ML Model, shows how to track model performance, detect data drift, monitor resource usage, and conduct controlled rollouts. Chapter 8, Notification and Alerting in MLOps, teaches you how to use event-driven alerts (e.g., via Event Grid) to detect anomalies and trigger automated responses. Chapter 9, Automating the ML Lifecycle with ML Pipelines and GitHub Workflows, details how to orchestrate model deployment using GitHub Actions and infrastructure-as-code practices. Chapter 10, Using Models in Real-world Applications, presents three cloud-based case studies (Azure, GCP, AWS) to demonstrate MLOps in practical industry settings. Chapter 11, Exploring Next-Gen MLOps, introduces LLMOps, showing how to work with large language models (LLMs), Retrieval-Augmented Generation (RAG), and responsible AI practices.