Nehodí sa? Žiadny problém! Tovar môžete vrátiť až do 30 dní
S darčekovým poukazom nešliapnete vedľa. Obdarovaný si za darčekový poukaz môže vybrať čokoľvek z našej ponuky.
Až 30 dní na vrátenie tovaru
Deploying a machine-learning model is easy to demonstrate. Building a machine-learning system that you can evaluate, release, operate, troubleshoot and recover with confidence is a much harder engineering problem.
Deploying Machine Learning with Amazon SageMaker AI teaches you how to transform a machine-learning workload from a defined business requirement into an evidence-driven production system on AWS.
Using a practical customer-churn workload, you will build a complete engineering path from data preparation and managed training through evaluation, controlled release, deployment, automation and production operations.
You will learn to distinguish between four fundamental measures of success: compute success, model quality, deployment health and continued predictive usefulness.
What you will learn
Data and training: Prepare and validate data, run SageMaker Processing and Training Jobs, and verify that managed computation produces the expected outputs and model artefacts.
Evaluation and controlled release: Evaluate model behaviour using precision and recall requirements, preserve decision thresholds, apply evidence-based release gates, and register and approve model versions.
Deployment and automation: Deploy models for real-time inference, verify prediction behaviour, automate machine-learning workflows with SageMaker Pipelines, and integrate controlled CI/CD release processes.
Infrastructure and security: Manage infrastructure with Terraform, understand AWS identity and execution-role boundaries, establish resource ownership, and control operational costs.
Production operations: Collect operational evidence, distinguish service health from model quality and data drift, diagnose failures systematically, recover workloads, and verify recovery and cleanup.
Throughout the book, one engineering method connects these capabilities:
UNDERSTAND → BUILD → DEPLOY → AUTOMATE → OBSERVE → DIAGNOSE → RECOVER
The emphasis is on evidence at every stage: what should happen, what actually happened, whether the result satisfies the requirement, and where to investigate when expected and observed behaviour diverge.
Who this book is for
DevOps engineers, cloud and platform engineers, software developers, solutions architects and technical professionals seeking a practical route into production machine-learning engineering on AWS.
No prior experience with statistical modelling or machine learning is required.
Companion source code
The official GitHub repository provides ready-to-use source files, configurations, installation guidance and execution instructions organised by chapter:
https://github.com/vebai/sagemaker-ai-companion
The book remains fully self-contained, providing the engineering concepts, architectural decisions, implementation sequence, verification methods and troubleshooting discipline behind each stage.
Develop the engineering judgement and practical capability to build, release, operate and recover production machine-learning systems on AWS.
Ahoj! Som Libroamiko, tvoj knižný radca.
Ako ti môžem pomôcť?