Become a Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-901/AI-103)

AI-901: Azure AI fundamentals (1 day)

Module 1: identify AI concepts and responsibilities

• Responsible AI principles:
  • fairness and reliability;
  • data privacy and security;
  • inclusiveness, transparency and accountability.
• AI model components and configurations:
  • how generative AI models work;
  • selecting an appropriate model: LLM, small language model or multimodal model;
  • deployment options and configuration settings.
• Common AI workloads:
  • generative and agentic AI;
  • text analysis and speech recognition;
  • computer vision;
  • information extraction.

Module 2: implement AI solutions with Microsoft Foundry

• Exploring the Microsoft Foundry portal and SDK.
• Deploying generative and multimodal models in Foundry.
• Creating a lightweight client application: chat, text analysis, vision or information extraction.
• Creating and testing a first agent with Foundry tools.

Lab

• Deploy a model in Microsoft Foundry and create a simple chat application.

AI-103: develop AI applications and agents on Azure (4 days)

Module 1: plan and manage an Azure AI solution

• Selecting the appropriate Foundry model and service for each task:
  • large language models (LLMs) and small language models;
  • multimodal models and Foundry tools.
• Designing Azure infrastructure for AI applications and agents:
  • selecting retrieval and indexing methods;
  • integrating memory, tools and knowledge for agentic solutions.
• Securing an Azure AI solution:
  • configuring managed identity and private networking;
  • configuring safety filters, guardrails and risk detection.

Lab

• Plan and secure an Azure AI solution architecture.

Module 2: implement generative and agentic AI solutions

• Designing retrieval-augmented generation (RAG) solutions:
  • grounding, citations and source traceability;
  • vector and hybrid search.
• Designing agentic workflows:
  • function calling and tool integration;
  • memory and approvals in an agent workflow;
  • multi-agent orchestration;
  • agent monitoring and tracing.
• Prompt engineering to optimise responses.

Lab

• Build an agent with RAG, function calling and vector search using the Foundry SDK.

Module 3: implement computer vision solutions

• Image analysis: classification, object detection and text extraction.
• Training and publishing a custom vision model.
• Analysing videos and live streams.

Lab

• Implement a computer vision solution with a Foundry model.

Module 4: implement text analysis and speech solutions

• Text analysis: key phrase extraction, sentiment analysis and personally identifiable information (PII) detection.
• Translating text and documents.
• Speech processing:
  • speech synthesis and recognition;
  • integrating voice as an agent modality.

Lab

• Build a speech-to-text and text-to-speech flow for agentic interaction.

Module 5: implement information extraction solutions

• Information extraction through multimodal pipelines: OCR, layout analysis and field extraction.
• Using Content Understanding to produce representations usable by agents and RAG solutions.
• Creating analysers that generate structured outputs for downstream reasoning.

Lab

• Extract and structure document data with Content Understanding for use by an agent.