Opleiding: Python: Building AI Agents with LangChain (English) (Virtual)

Lesmethode :

Virtueel

Algemeen :

AI agents combine a language model with tools, allowing an application to reason about a task, choose suitable actions and work step by step towards a result. Unlike a chatbot, an agent can call Python functions, retrieve information, process data and interact with external systems. In this two-day, hands-on course, you learn how to build AI agents in Python with LangChain. You learn the difference between a language-model application and an AI agent and work with the agent loop of model, tool and result.

You learn how LangChain's create_agent assembles this agent loop on top of the LangGraph runtime. You connect a language model, turn Python functions into tools with the @tool decorator, and define clear tool names, descriptions and input schemas so an agent can choose between multiple tools. You also work with system prompts, messages and agent state for short-term memory and structured output. In addition, you learn how to handle failing tools and invalid results, apply stop conditions and iteration limits, and use tracing to inspect agent decisions and tool calls.

During the course, you spend most of your time writing code, with explanations kept short. You invoke agents, process their results and finish by building a working end-to-end agent that uses multiple tools to carry out a practical task. This gives you direct experience building agentic AI applications in Python.

Doel :

After this course, you can build tool-using AI agents in Python with LangChain, including tools, prompts, conversation state and structured output. You can also handle failures, apply execution limits, trace agent decisions and build an end-to-end agent application. The course focuses on building agents in Python; training or fine-tuning language models is not covered.

Doelgroep :

Python developers, software developers, data and AI engineers, automation engineers and other technical professionals who want a practical introduction to building AI agents.

Voorkennis :

The following prior knowledge is required:

- Leren programmeren in Python

or

- De programmeertaal Python voor ervaren programmeurs

- Basic understanding of APIs and JSON

Onderwerpen :

- Agents and the agent loop
+nbsp;+nbsp;+nbsp;- Introduction to AI agents
+nbsp;+nbsp;+nbsp;- Agents versus chatbots and workflows
+nbsp;+nbsp;+nbsp;- The agent execution loop
+nbsp;+nbsp;+nbsp;- LangChain agent architecture
- Creating an agent with LangChain
+nbsp;+nbsp;+nbsp;- Language models and provider configuration
+nbsp;+nbsp;+nbsp;- Creating an agent with create_agent
+nbsp;+nbsp;+nbsp;- System prompts and agent instructions
- Building and describing tools
+nbsp;+nbsp;+nbsp;- Creating tools from Python functions
+nbsp;+nbsp;+nbsp;- The LangChain @tool decorator
+nbsp;+nbsp;+nbsp;- Tool names, descriptions and input schemas
+nbsp;+nbsp;+nbsp;- Multiple tools and tool selection
- Running an agent and managing state
+nbsp;+nbsp;+nbsp;- Invoking an agent
+nbsp;+nbsp;+nbsp;- Messages and agent state
+nbsp;+nbsp;+nbsp;- Short-term memory
+nbsp;+nbsp;+nbsp;- Structured output
- Reliability and control
+nbsp;+nbsp;+nbsp;- Error handling and tool failures
+nbsp;+nbsp;+nbsp;- Stop conditions and iteration limits
+nbsp;+nbsp;+nbsp;- Tracing and debugging agent execution
- Putting it all together
+nbsp;+nbsp;+nbsp;- Building an end-to-end AI agent