Building AI Agents with Microsoft Azure

via Udacity

Go to class Write review

Details

Go to class

Provider

Udacity
Pricing

Paid Course
Languages

English
Certificate

Certificate Available
Duration & workload

15 hours
Sessions

On-Demand

Found in

Overview

This course explores the intricacies of constructing AI agents utilizing Microsoft Azure. You will embark on a journey through the fundamentals of agent design before mastering tools that extend agent capabilities. Key focus areas include the implementation of structured outputs with Pydantic, state management, and memory systems, both short- and long-term. The curriculum emphasizes integration with external tools and databases, particularly through Cosmos DB and Bing Search to enhance functionality. Additionally, you will explore evaluation techniques for agents. By the end of the course, you will build a sophisticated AI travel concierge agent, contextualizing their learning in a hands-on project.

Syllabus

Introduction to Building AI Agents with Microsoft Azure

Get to know your course instructors, set up OpenAI resources, and get an overview of the course.

Extending Agents with Tools

Extend AI agents beyond text with tool integrations, enabling reliable real-time actions and data access.

Building Agents with Semantic Kernel and Tools

Learn to empower AI agents in Python by building and registering custom tools with Semantic Kernel, allowing access to real-time external data and enhanced functionality.

Structured Outputs

Discover structured outputs in AI: transform responses into actionable JSON for integration. Utilize schemas, parsers, and function calls to enhance reliability and automation in workflows.

Implementing Structured Outputs using Pydantic and Semantic Kernel

Learn to compel LLMs to output structured, validated JSON by combining Pydantic models, structured prompts, and Semantic Kernel's automatic tool invocation.

Agent State Management

Explore agent state management with state machines. Learn how agents track user input, instructions, and tool use for complex workflows, ensuring adaptability and reliability.

Implementing Agent State Management using Semantic Kernel

Learn to build robust AI agents using a Python finite state machine to define phases, manage agent state, guide LLMs, and ensure predictable, auditable multi-step workflows.

Short-Term Agent Memory

Explore short-term memory in AI agents, enhancing coherence via state, ephemeral, and ephemeral memory strategies for efficient context retention in active sessions.

Adding Short-Term Memory to an Agent using Semantic Kernel

Learn to add short-term memory to an agent using Python, enabling it to track conversation history, manage memory limits, and handle follow-up questions in dialogues.

External Tools and APIs

Explore using external APIs for real-time data, dynamic actions, and authenticating agents. Discover MCP, a protocol standardizing AI’s tool interoperability and safety.

Integrating External Tools and APIs using Semantic Kernel

Learn to enhance LLMs with real-time data using Semantic Kernel, integrating external APIs as plugins, automatic tool selection, and maintaining conversation context for expert agents.

Web Search Agents

Equip agents to search web for real-time, unstructured info. Ground responses in evidence using APIs, handle noise, and avoid hallucination for credible answers.

Creating Web Search Agents using Bing Search

Build AI agents with real-time Bing web search using Semantic Kernel, Azure AI Agents, prompt engineering for JSON, and auto tool selection in Python.

Interacting with Databases

Equip agents to access and modify structured data by using SQL for interaction and vector databases for semantic tasks, ensuring seamless integration with private systems.

Building RAG Systems using Cosmos DB

Learn to build AI agents with Retrieval-Augmented Generation using Azure Cosmos DB, including data ingestion, text-based retrieval, and robust agent integration.

Agentic Retrieval Augmented Generation

Discover Agentic RAG: Enhance RAG by enabling reflection, query reformulation, and intelligent adaption for nuanced answers. Master retrieval, reasoning, and retry loops.

Implementing Agentic RAG using Cosmos DB

Build agentic RAG with Cosmos DB: create a self-correcting AI agent that refines queries, uses LLMs to assess document quality, and leverages hybrid search with RRF for better answers.

Long-Term Agent Memory

Explore long-term agent memory: understand semantic, episodic, and procedural memories. Learn storage strategies and best practices for personalized, coherent interactions.

Maintaining Long-Term Agent Memory using Cosmo DB

Learn to build AI agents in Python with long-term memory, enabling them to recall past interactions for context-aware, personalized conversations using memory storage and retrieval.

Agent Evaluation

Agent Evaluation guides assessing an agent’s task completion, quality, tool use, and system metrics using response, step, or trajectory strategies to ensure reliable and efficient operations.

Implementing Agent Evaluation using Semantic Kernel

Learn to evaluate AI agents by combining rule-based structural checks with LLM-as-judge semantic analysis for comprehensive quality assessment.

AI Travel Concierge Agent

Build a Python-based travel concierge agent for a premium bank that uses tools, stateful orchestration, RAG, and dual memory to plan trips end-to-end and recommend the best credit card.