Skip to content

Plan and create an Azure AI solution

🧩 Problem Statement

You need to develop comprehensive AI solutions that combine multiple capabilities: - Integrate machine learning models with AI services - Build applications using generative AI and prompt engineering - Create scalable solutions that handle various AI tasks - Ensure responsible AI implementation

Key Requirements:

  • Choose appropriate AI services for specific capabilities
  • Manage resources efficiently (single vs multi-service)
  • Select proper development tools and SDKs
  • Implement responsible AI principles
  • Handle regional availability and cost considerations

💡 Solution with Azure

Azure AI Services provides a comprehensive suite of pre-built AI capabilities that developers can integrate into applications without deep machine learning expertise. Combined with Azure AI Foundry, you get a complete platform for AI development.

🧩 Required Components

Core AI Capabilities

  • Generative AI: Generate original responses to natural language prompts
  • Agents: Execute tasks autonomously (e.g., executive assistants, meeting schedulers)
  • Computer Vision: Process visual input from images, videos, and live camera streams
  • Speech: Recognize/synthesize speech, enable voice interactions
  • Natural Language Processing: Process written/spoken language, analyze sentiment
  • Information Extraction: Extract key information from documents, forms, images
  • Decision Support: Make predictions for business decision making

Azure AI Services

  • Azure OpenAI: Access to GPT models for generative AI
  • Azure AI Vision: Computer vision capabilities with APIs
  • Azure AI Speech: Text-to-speech and speech-to-text transformation
  • Azure AI Language: NLP capabilities including entity extraction, sentiment analysis
  • Azure AI Foundry Content Safety: Advanced algorithms for processing offensive content
  • Azure AI Translator: State-of-the-art language translation
  • Azure AI Face: Detect, analyze, and recognize human faces
  • Azure AI Custom Vision: Train and use custom vision models
  • Azure AI Document Intelligence: Extract fields from documents
  • Azure AI Content Understanding: Multi-modal content analysis
  • Azure AI Search: Create searchable indexes with AI skills

🛠 Architecture & Development

🔹 Resource Management

Single-Service Resources - Create standalone resources for specific services - Best for applications using limited AI capabilities - Examples: Azure AI Vision, Azure AI Language

Multi-Service Resources - Encapsulates multiple services in a single resource - Includes: OpenAI, Speech, Vision, Language, Foundry Content Safety, Translator, Document Intelligence, Content Understanding - Easier management for applications using multiple capabilities - Single endpoint and authorization key

🔹 Azure AI Foundry

Platform Benefits - Centralized project organization and resource management - Web-based portal for visual interface - Azure AI Foundry SDK for programmatic access

Project Types

  1. Foundry Projects
  2. Associated with Azure AI Foundry resource
  3. Support for deploying models (OpenAI, Azure AI Foundry Agent Service, Azure AI services)
  4. Ideal for generative AI chat apps and agents

  5. Minimal administrative resource management

  6. Hub-based Projects

  7. Associated with Azure AI hub resource
  8. Support for Prompt Flow development
  9. Connected Azure storage and Key vault resources
  10. Advanced scenarios like fine-tuning models
  11. Better for collaborative projects with data scientists and ML specialists

🔹 Development Tools & SDKs

Development Environments - Microsoft Visual Studio - VS Code (with Azure AI Foundry VS Code container image) - GitHub integration with GitHub Copilot

SDKs and APIs - Azure AI Foundry SDK: Connect to projects and access resource connections - Azure AI Services SDKs: Service-specific libraries for multiple languages (C#, Python, Node.js, Java) - Azure AI Foundry Agent Service: Integrate with frameworks like AutoGen and Semantic Kernel - Prompt Flow SDK: Implement orchestration logic for generative AI

VS Code Container Image Benefits - Pre-installed SDK packages - Hosted web application in browser - Latest versions of required tools - Compute resources scalable to project needs

🆚 Single vs Multi-Service Resources

Aspect Single-Service Multi-Service
Use Case Limited AI capabilities needed Multiple AI capabilities required
Management Individual resource per service Single resource for all services
Cost Pay per service used Consolidated billing
Complexity Simple for single capability Simplified for multi-capability apps
Authorization Separate keys per service Single endpoint and key

🧠 Best Practices & Considerations

Responsible AI Principles

  1. Fairness
  2. Treat all people fairly regardless of demographics
  3. Review training data for bias

  4. Evaluate performance across user populations

  5. Reliability and Safety

  6. Rigorous testing and deployment management
  7. Account for probabilistic nature of ML models

  8. Apply appropriate thresholds for confidence scores

  9. Privacy and Security

  10. Protect personal data in training and production
  11. Implement appropriate safeguards

  12. Respect user privacy expectations

  13. Inclusiveness

  14. Design for diverse user groups
  15. Test with varied input sources

  16. Ensure accessibility features

  17. Transparency

  18. Make users aware of AI system usage
  19. Share confidence scores and limitations

  20. Clear data usage and retention policies

  21. Accountability

  22. Define governance framework
  23. Clear responsibility assignment
  24. Regular review of system performance

Cost and Regional Considerations

  • Check service availability in target regions
  • Use Azure pricing calculator for cost estimation
  • Consider usage patterns for pricing model selection
  • Monitor actual usage against estimates

Development Best Practices

  • Start with Azure AI Foundry for simplified management
  • Use multi-service resources for complex applications
  • Leverage VS Code container image for consistent development
  • Implement proper error handling for AI service calls
  • Use appropriate SDKs for your programming language

🎯 Exam Simulation Questions

Q: Which Azure resource provides language and vision services from a single endpoint? ✅ Azure AI Services (multi-service resource)

Q: You plan to create a simple chat app that uses a generative AI model. What kind of project should you create? ✅ Azure AI Foundry project

Q: Which SDK enables you to connect to resources in a project? ✅ Azure AI Foundry SDK

Q: What is a key principle of responsible AI regarding system transparency? ✅ Users should be made aware of the AI system's purpose, limitations, and confidence scores

Q: Which development tool provides pre-installed SDK packages for Azure AI development? ✅ Azure AI Foundry VS Code container image

Q: What capability would you use to automatically generate property descriptions for real estate listings? ✅ Generative AI (using Azure OpenAI)

Q: Which service combination is included in a multi-service Azure AI Services resource? ✅ OpenAI, Speech, Vision, Language, Foundry Content Safety, Translator, Document Intelligence, Content Understanding