AI Agent Development with LangChain and Azure

Packt via Coursera

Go to class Write review

Overview

Google, IBM & Meta Certificates – 40% Off

One Coursera Plus subscription covers most Professional Certificates on Coursera.

Unlock All Certificates

Develop practical AI agent skills by building LangChain and LangGraph workflows on Azure. The course covers ReAct agents, tools, MCP integrations, sandboxed code execution, memory, human oversight, multi-agent orchestration, and LangSmith observability. AI agents are reshaping how software systems use large language models by combining reasoning with actions, tools, memory, and external systems. For developers and cloud practitioners, this skill set is increasingly relevant as organizations look for agentic applications that can retrieve information, execute tasks, and remain observable and controllable. The learning path begins with LangChain fundamentals and Azure AI Foundry model deployment using Terraform and secure configuration. It then moves into ReAct agents, custom tools, function calling, Model Context Protocol servers for Microsoft Learn and web search, sandboxed Python and shell execution, Azure Cosmos DB memory, human-in-the-loop oversight, middleware hooks, sub-agent orchestration, and LangSmith tracing. Practical demonstrations connect each concept to working implementations, helping you move from prompt-driven prototypes to structured agent workflows. You will gain experience designing agents that call tools, preserve conversational context, involve humans when needed, and support debugging through observability. By the end of this course, you will be able to design, build, and operate Azure-based AI agents for real-world development scenarios. This course is designed for Python developers, Azure engineers, cloud architects, and AI/ML practitioners will benefit from this course, especially those moving from prompts to agentic systems. Basic Python is expected; REST APIs, JSON, Azure access, and VS Code or Jupyter experience are helpful. Concepts are introduced through short explanations before moving into implementation. Demonstrations connect LangChain, LangGraph, Azure AI Foundry, MCP servers, and supporting services in a progressive workflow. Learners build from a first ReAct agent to tool use, memory, human oversight, orchestration, and observability. This course is based on Modern AI Agent Development Using LangChain and Microsoft Azure, by Houssem Dellai. This video is licensed and distributed by Packt. All rights reserved. Packt is one of the world's most prolific publishers of cutting-edge technical content. For over two decades we've made it our mission to curate and publish the knowledge of only the very best technical experts. We focus on real-world courses that help our customers get the job done, with coverage that extends across a wide range of established and cutting-edge technical topics. If you're an individual or an organisation that embraces learning by doing, Packt is the perfect fit for you.

Syllabus

Introduction
- This module introduces the importance of AI agents, their role in extending LLMs, and how they integrate tools, memory, human oversight, and MCP servers to enable action-oriented AI systems.
LangChain Fundamentals
- This module provides an introduction to LangChain fundamentals, covering the setup of Azure AI Foundry models using Terraform, the creation of ReAct agents, and the deployment of language model agents. Learners will gain hands-on experience with real-world prompts and secure configurations.
Tools & Function Calling
- This module explores how tools and function calling enhance AI agent capabilities, allowing them to perform task-oriented actions. Learners will gain an understanding of integrating custom tools, using ReAct agents, and implementing function-based workflows. It provides practical insights into connecting tools to language models for real-world applications.
MCP Server - Microsoft Learn Docs
- This module explores the Model Context Protocol (MCP) and its role in enabling AI agents to interact with external systems like Microsoft Learn. Learners will gain an understanding of how to connect agents to an MCP server and use structured tool calls to access documentation. The module also covers key concepts such as tool discovery and server setup.
MCP Server - Self-hosted Web Search
- This module covers the setup, integration, and operational principles of self-hosted MCP servers for real-time web search. Learners will explore how to connect AI agents to external knowledge sources, implement error handling, and develop custom tools for seamless web interaction.
Dynamic Sessions — Python & Shell
- This module covers the use of dynamic sessions for executing Python and shell commands in safe, isolated environments. Learners will explore how to configure and use these sessions within Azure Container Apps to enhance security and control over code execution.
AI Agent Memory with Azure Cosmos DB
- This module explores how AI agents can retain and manage conversational context using Azure Cosmos DB. Learners will understand the role of short-term memory, implement memory storage techniques, and analyze strategies for handling long conversations effectively.
Human-in-the-Loop (HITL)
- This module explores the role of human oversight in AI agent workflows, focusing on how to implement and use human-in-the-loop middleware to ensure safety, control, and reliability in automated systems. Learners will gain practical knowledge on reviewing, editing, and intervening in AI-generated actions through real-world examples and scenarios.
Before & After Model Hooks
- This module explores how to use before and after model hooks to customize and control AI agent behavior during model execution. Learners will gain practical skills in implementing middleware hooks for inspection, modification, and security in agent workflows. The content provides a foundation for managing agent interactions effectively.
Sub-Agents & Orchestration
- This module explores the design and implementation of multi-agent systems, focusing on how specialized agents collaborate and delegate tasks through orchestration patterns. Learners will gain an understanding of context management, performance optimization, and the use of MCP-enabled tools for scalable AI solutions.
Logging & Observability with LangSmith
- This module introduces the principles of logging and observability in AI agents, focusing on how LangSmith tracing enables effective monitoring, debugging, and optimization of agent-driven workflows. Learners will gain insights into identifying errors, analyzing performance, and improving system efficiency.
Outro
- This module provides a comprehensive summary of the agent-building process and outlines practical steps for applying the course concepts in real-world scenarios. Learners will gain clarity on the key takeaways and how to continue their learning journey.

Taught by

Packt - Course Instructors

Reviews

Start your review of AI Agent Development with LangChain and Azure

Go to class

Taught by

Never Stop Learning.