| Author/Contributor(s): | Falk, Kim |
| Publisher: | Manning |
| Date: | 5/25/2027 |
| Binding: | Paperback |
| Condition: | NEW |
Customers expect fast, accurate suggestions that are perfectly aligned to their preferences, even when they have limited historical data and provide minimal input. This book gives you a guided tour to the new era of recommenders, powered by LLMs, agents, and context engineering. Author Kim Falk takes you from your first introduction to recommenders all the way to evaluation and production-ready systems. You'll explore established and cutting-edge frameworks and algorithms as you learn how to customize and improve their outputs to fit your specific needs.
Each recommender system is a unique blend of data engineering, performance optimization, and, now, also AI-led tools like agents and RAG. As a developer, you also need an efficient process to interpret the domain in which you’re working and to train your system to respond to actual customer demand. And you have to understand your algorithms—where they’re strong, where they’re blind, how to shape output to the constraints your business cares about, and how to evolve with changing needs.
AI Recommender Systems introduces the complete pipeline of components that make up a modern recommender system, from candidate generation and filtering to scoring and reranking, and describes where LLMs can help. You'll learn to read the signals hiding in your data, including content metadata and user behavior, and to evaluate recommenders against the specific use cases they need to support. Along the way, you’ll learn how to integrate custom prompts, retrieval-augmented generation, and agents into new designs as well as how to retrofit them into your existing systems.
Finally, because most recommender systems are customer-facing, performance is a key aspect of their success. You’ll learn how to handle the cold-start problems every real system eventually hits and how to put a recommender into production with the monitoring it needs.
What's inside
• Build an AI-first recommender system from the ground up
• Evaluate recommenders against real business use cases
• Assemble candidate, filtering, scoring, and reranking pipelines
• Deploy and monitor recommenders in production
About the reader
For data workers, developers, and researchers building and deploying recommender systems. Readers need basic Python, statistics, and machine learning.
About the author
Kim Falk has worked with recommender systems for a decade and a half, currently as a principal recommender engineer. He has designed and implemented recommender systems serving millions of users across domains including news, media, and e-commerce. He is the author of Practical Recommender Systems. A software engineer by training, he cares about making research runnable in production.