Building agents the LangGraph way!
1. Building agents the LangGraph way!
Hi, and welcome to this course on building multi-agent systems with LangGraph. I'm James, and I'll be your instructor for this course.2. What you'll build...
In this course, you'll create an agentic assistant to help retrieve and summarize information on Fortune 500 companies. This assistant will be able to retrieve summary information about the company from the internet, stock performance data from CSV files, and run Python code to generate stock performance visualizations.3. Agents: the LangGraph way!
We'll be using the LangGraph framework to build these agents. LangGraph is part of the LangChain ecosystem, specifically designed for building and orchestrating production-ready agents. These agents are constructed as nodes and edges, which we'll see in a few moments, and this approach provides a high-degree of control and customizability over the workflow. Let's briefly zoom out and discuss the need for AI agents.4. Limitations of LLMs
Recall that LLMs have a knowledge cut-off: they only know about information or events included in their training data, and they can only typically generate text, or sometimes images if they have multi-modal capabilities. Because it cannot interact with the outside world, an LLM on its own is not an agent. To perform actions in the real-world, and hence gain agency and become "agentic", LLMs require a means to trigger actions, which they can do by calling tools.5. Agents: the story from the inside
This is the essence of an agent. A prompt comes in from a user to complete a task,6. Agents: the story from the inside
it goes to the LLM, which can choose to select an appropriate tool to complete the task.7. Agents: the story from the inside
These tools can interact with the real-world, such as pulling real-time data from databases, APIs, documents, or search engines, triggering events in third-party applications like Google Calendar, or running code to perform calculations or analyze data. The capabilities of tools are really only bounded by what can be programmed into a function.8. Agents: the story from the inside
The orchestration layer maps out how these interactions take place, and handles things like memory. We'll be using LangGraph for the orchestration.9. What you'll build...
Returning to our company research assistant, we can express this in terms of nodes and edges, which we'll replicate in the LangGraph code we'll write in the exercises.10. Nodes and edges
Nodes indicate a component in the workflow, and edges represent paths in which messages can be exchanged. A node is built for the LLM, and another for the tools it has access to. Edges are added to define where information should travel and depending on what conditions!11. Course format
Throughout the course, we'll be working side-by-side to tackle exercises, which will be completed in DataLab Notebooks with accompanying videos. DataLab is DataCamp's proprietary cloud-based notebook IDE. If you've used other notebook IDEs, you should find DataLab pretty intuitive; if you're not familiar with notebooks, you can think of a notebook as a series of cells to execute code in. If you run one cell, any variables or libraries defined in that cell will be available in subsequent cells, and so on. In the exercise, you can either start the video and code-along with us, or you can watch the video first and code afterwards - the choice is yours!12. Course format
You can find detailed information to help you complete the tasks in the instructions section, and in the notebook.13. Let's practice!
Time to get familiar coding agents with LangGraph!Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.