Building Agentic RAG Systems with LlamaIndex

Static retrieval-augmented generation (RAG) is often not enough when dealing with complex, multi-step data queries. To build truly smart applications, you need autonomous agents that can plan, reason, and make decisions about which data sources to access. This text-based course guides you from the fundamental concepts of agentic architectures to building robust, self-correcting RAG systems using LlamaIndex. You will transition from basic document retrieval to constructing agents capable of complex reasoning, multi-document summarization, and intelligent debugging. What you'll learn: - Understand the core terminology of Agentic RAG, including the transition from standard retrieval to autonomous reasoning loops. - Configure LlamaIndex query engines and tools to allow agents to interact dynamically with your data. - Build routing agents that intelligently select the best data source or retrieval strategy for a given user query. - Implement multi-step reasoning patterns to solve complex, multi-document Q&A and summarization tasks. - Apply debugging and evaluation techniques to monitor agent decisions and optimize retrieval accuracy. - Explore modern agentic workflows and tool-calling patterns to ensure reliable, production-ready performance. You will start with foundational definitions and basic setup before moving step-by-step through tool integration, routing logic, and advanced agentic workflows. Each concept is reinforced with clear, written code examples and practical scenarios designed to build your confidence. This course is designed for software developers, data enthusiasts, and AI beginners who want to move beyond simple search applications. A basic familiarity with Python is recommended, but no prior experience with LlamaIndex or agentic systems is required. Start reading today to unlock the power of autonomous, data-aware agents.