courseraАнализ данных

Advanced RAG Patterns

Курс от Coursera

Средний≈ 2 чАнглийский

О курсе

Advance RAG Patterns is an intermediate course designed for AI developers and ML engineers who have built a basic RAG pipeline but find it still fails on complex or nuanced queries. While foundational RAG reduces hallucinations, production-grade AI demands greater reliability, accuracy, and reasoning. This 2-hour course moves beyond the basics to teach you how to engineer robust, intelligent, and self-correcting systems. Focused on practical, job-ready skills, this course dives deep into cutting-edge architecture. You will learn to implement and evaluate a suite of advanced patterns, including Corrective RAG for query rewriting, Self-RAG for source validation, and Agentic RAG for multi-hop problem-solving. Through hands-on, in-browser projects, you will A/B test these different architectures, analyze their performance against key metrics, analyze different embedding services, and make data-driven decisions on improving accuracy. By the end, you'll be able to not just build, but architect and defend production-ready RAG systems that are both powerful and trustworthy.

Навыки, которые вы освоите

Retrieval-Augmented Generation Generative AI Agents Performance Tuning Generative AI Agentic systems Performance Testing Agentic Workflows Model Evaluation Verification And Validation Large Language Modeling Embeddings Prompt Engineering Prompt Patterns MLOps (Machine Learning Operations)

Программа курса

2 модулей · 10 учебных материалов

01Enhancing RAG with Correction and Validation Patterns5 материалов

AI Coach: "My RAG Bot Is Still Wrong Sometimes!"DIALOGUE The Theory of Self-Correction: Corrective vs. Self-RAGЧтение How to Implement a Self-RAG and a Corrective RAG LoopВидео Hands-On Learning: Add a Validation Step to a RAG Pipeline

Учитесь у экспертов

Professionals in the Industry

Преподаватель курса