Oreilly – LLMs in Production, Video Edition 2025-1

Updated August 10, 2026 2.4 GB

Download

About this item

Description

LLMs in Production, Video Edition. This course teaches you how to safely and efficiently deploy large language model-based applications in production and manage their lifecycle. With clear explanations and rich examples, this practical book shows how these models work, how to interact with them, and how to integrate them into applications. This course helps you understand how large language models differ from traditional software and other machine learning systems, learn best practices for real-world applications, and avoid common mistakes with expert advice. This course provides essential insights into implementing machine learning operations to seamlessly and easily transition a model to production. This course provides practical guidance on topics as diverse as preparing the right training dataset, building the platform, managing large model sizes, prompt engineering, retraining, load testing, controlling costs, and ensuring security. Given the high cost of building and difficulty of changing these models, careful planning, strong data standards, and calculated technical implementation are essential. Integrating these models into operational products directly impacts all aspects of the application, including the application lifecycle, data pipeline, computational costs, and security issues.

What you will learn

Introduction to LLM principles:
Understand the basic principles of LLMs and the technology behind them.
LLM Evaluation and Selection:
Assessing when to use a pre-built LLM and when to build your own LLM.
Scaling and training:
Efficiently scaling the ML platform to meet the needs of LLMs.
Training on fundamental LLM models and fine-tuning an existing LLM.
Advanced deployment and architectures:
Deploying LLMs in the cloud and edge devices using complex architectures such as PEFT and LoRA.
Building applications:
Building applications that leverage the strengths of LLMs and mitigate their weaknesses.
Operational Management:
Balancing cost and performance.
Retraining and load testing.
Optimization of models for commodity hardware.
Deploy on a Kubernetes cluster.

This course is suitable for people who:

Data scientists and ML engineers who are familiar with Python and cloud deployment fundamentals.
Engineering teams looking to develop an LLMOps application to seamlessly transition an AI application from design to delivery.
Professionals who want to learn techniques for LLM dataset preparation, low-cost training hacks like LORA and RLHF, and industry benchmarks for model evaluation.