Azure OpenAI Service Training is a practical, developer-focused program designed to help participants build and integrate generative AI solutions using Azure OpenAI Service. The course provides hands-on experience with large language models (LLMs), prompt engineering, AI-powered applications, and modern techniques for creating intelligent solutions on Microsoft Azure.
AI-050: Develop Generative AI Solutions with Azure OpenAI Service Course Overview
This course introduces developers and AI professionals to building generative AI solutions using Azure OpenAI Service and Microsoft Foundry. Participants learn how to select and deploy generative AI models, develop applications using APIs and SDKs, apply prompt-engineering techniques, ground model responses with organizational data, and implement responsible AI safeguards.
The course emphasizes the practical knowledge required to create secure, reliable, and context-aware generative AI applications on Microsoft Azure.
Duration 1 Days – 7 hrs.
Learning Objectives
- Explain the capabilities and common use cases of generative AI.
- Describe Azure OpenAI Service and its role within Microsoft Foundry.
- Create and configure an Azure OpenAI resource.
- Select and deploy an appropriate generative AI model.
- Use the Azure OpenAI API and supported SDKs in an application.
- Design effective system instructions and user prompts.
- Control model output through prompt parameters and structured instructions.
- Develop conversational generative AI applications.
- Generate and use vector embeddings.
- Ground model responses using proprietary or organizational data.
- Explain the fundamental architecture of retrieval-augmented generation.
- Apply responsible AI, content-safety, security, and monitoring practices.
Target Audience
- AI engineers
- Software developers
- Application developers
- Cloud developers
- Data scientists
- Machine-learning engineers
- Solutions architects
- Technical consultants
- IT professionals responsible for developing AI-enabled applications
Prerequisites
- Basic knowledge of artificial intelligence and machine-learning concepts.
- General familiarity with Microsoft Azure services and the Azure portal.
- Experience programming in Python or C#.
- Basic experience working with REST APIs and JSON.
- Familiarity with authentication, cloud resources, and application-development concepts.
- An active Azure subscription with appropriate access to Azure OpenAI Service or Microsoft Foundry resources.
Course Outline
Day 1
Module 1: Introduction to Generative AI
- Generative AI concepts and terminology
- Large language models and foundation models
- Tokens, prompts, completions, and context windows
- Common generative AI workloads and business scenarios
- Generative AI capabilities and limitations
Module 2: Azure OpenAI Service and Microsoft Foundry
- Azure OpenAI Service overview
- Microsoft Foundry projects, tools, and resources
- Available model families and model-selection considerations
- Creating and configuring Azure AI resources
- Deploying and testing a generative AI model
- Understanding quotas, availability, and access controls
Module 3: Developing Applications with Azure OpenAI
- Azure OpenAI endpoints and authentication
- Using REST APIs and supported SDKs
- Configuring model requests and parameters
- Processing and presenting model responses
- Managing conversational context
- Handling errors, rate limits, and service responses
Module 4: Prompt Engineering
- Principles of effective prompt design
- System instructions, user prompts, and assistant messages
- Zero-shot and few-shot prompting
- Providing context and examples
- Controlling tone, format, and response structure
- Improving prompt reliability and consistency
- Mitigating prompt injection and unintended behavior
Module 5: Embeddings and Retrieval-Augmented Generation
- Understanding vector embeddings
- Semantic similarity and vector search
- Introduction to Azure AI Search
- Chunking, indexing, and retrieving source content
- Grounding responses with organizational data
- Retrieval-augmented generation architecture
- Adding citations and source references to responses
Module 6: Responsible and Production-Ready Generative AI
- Microsoft responsible AI principles
- Content filtering and content-safety controls
- Protecting sensitive and confidential information
- Security, identity, and access considerations
- Monitoring model usage and application performance
- Managing cost, scalability, and reliability
- Review of recommended production practices
