Welcome to Event-Driven Agentic Document Workflows built in partnership with LlamaIndex. I'm delighted that the instructor for this course is Laurie Voss, whose VP of developer relations at LlamaIndex. Thanks, Andrew. I'm jazzed to be back here with you and teach this course. Agentic document workflows are are agent-based applications that you can use to automate an end-to-end document processing workflow. While RAG systems answers simple questions about your data, agentic workflows can be built on top of RAG to help you process input documents in more sophisticated ways. In the architecture you learn about, an agent will identify key information that you will need to carry the tasks and then retrieve relevant materials using RAG. And finally, combine collected information into a structured output. For example, say you want to review a contract for compliance with certain regulations. The agent could parse the contracts, extract the key clauses, and match them with relevant clauses from a knowledge base of regulatory requirements, and finally generate a compliance summary. As another example, say you want to enrich a set of invoices with standardized product information. The agent can extract the item descriptions from the invoice, match them with the closest product code from the catalog using RAG, and it append the standardized information to that invoice. In this course, you apply this type of workflow to a practical application in which you build an agent that use the resume to follow the job application form. Laurie will walk you through how to build this from scratch using LlamaIndex's workflow abstraction, which is a good way to build event-driven systems, which is a key design pattern for building efficient sets of agents. That's right. LlamaIndex's workflows is an event-driven architecture that you will use to create your agent. You will encapsulate the agent's logic in a chain of steps, where each step emits events to trigger further steps. You will learn how you can use code branching and looping within the workflow. Create concurrent events, and collect multiple events at a given step. You'll apply these concepts to build your form filling agent step by step. You'll start by setting up your agent's RAG capability to parse the given resume, load it into a vector store, and create a query engine. Then you'll get your agent to parse the job application form, convert the blank spaces into a series of questions, and send them to the RAG pipeline. You'll then provide feedback to your agent and iterate together on the returned answers. You will communicate your feedback through text and then using your voice. Many people have worked to create this course. I'd like to thank Logan Markewich from LlamaIndex, who helped a great deal, and from DeepLearning.AI, Hawraa Salami also contributed to this course. Event-driven workflows are really important design pattern, and I see more and more businesses using it to design large language model-powered applications. I think you enjoy learning about these concepts. So, please go on to the next video and let's get started.