Learn AI Agent Architectures – Chaining and Routing
In this hands-on workshop, you’ll learn to design and implement AI agent architectures that enable structured decision-making, task automation, and workflow optimization. AI agents can process inputs, make intelligent decisions, and take actions autonomously. By leveraging Langbase’s composable agent architecture, you will build scalable and efficient AI systems without the need for complex frameworks.
AI agent architectures define how agents interact with Large Language Models (LLMs), tools, memory, and workflows to achieve specific objectives. These architectures range from simple single-step models to complex multi-agent systems capable of reasoning, routing, and executing multi-step processes.
This session will focus on two key architectures used for structuring AI workflows:
Prompt Chaining – A method where tasks are broken into sequential steps, with each LLM call using the previous step's result. This approach improves accuracy and control, making it ideal for structured processes like content generation, verification, and multi-step reasoning.
Agentic Routing – A system that classifies inputs and directs them to specialized AI agents for improved efficiency. Routing ensures that different tasks—such as summarization, reasoning, or coding—are handled by the most suitable AI model, optimizing both performance and cost.
You will learn to:
Understand AI agents and their architectures
Implement prompt chaining for structured workflows
Use agentic routing to optimize AI decision-making
Define workflows that integrate multiple AI agents
Deploy and scale AI agents using Langbase
Everyone is welcome to join this event, whether you're an experienced developer or an aspiring AI engineer—no matter your background. Feel free to share the invite link with others.
To get the most out of the session, make sure you have a code editor and Node v20+ installed beforehand. You should also be familiar with basic JavaScript and using the command line.