Develop AI apps and agents on Azure (AI-103)

1. Plan and Prepare to Develop AI Solutions on Azure
· Introduction to AI Applications and Agents: Overview of modern AI architectures.
· Microsoft Foundry Tools: Leveraging AI Foundry for managing model deployments and lifecycle.
· Responsible AI: Implementing guardrails, safety filters, and managing risk.
· Developer Tools & SDKs: Setting up Python, VS Code, and SDKs for agent development.

2. Develop Generative AI Apps in Azure
· Model Selection and Deployment: Choosing foundational models (LLMs, small language models) and deploying to endpoints.
· Prompt Engineering: Designing effective prompts to guide AI reasoning and outputs.
· Chat App Development: Using the ChatCompletions API to build conversational apps.
· Retrieval Augmented Generation (RAG): Grounding models with enterprise data for accurate answers.
· Optimization: Evaluating models using metrics such as groundness, relevance, and fluency.

3. Develop AI Agents on Azure
· Agent Architecture: Transitioning from chat to agents that break down tasks, reason, and act.
· Tool Integration: Connecting agents to APIs and external services (e.g., custom tools, plugins).
· Agent State and Memory: Managing context across long-running conversations.
· Foundry IQ: Building knowledge-enhanced agents using Foundry IQ for data connectivity.
· Multi-Agent Orchestration: Using Microsoft Agent Framework for multi-agent collaboration.
· Integration with Microsoft 365: Publishing agents to Teams and accessing M365 data.

4. Develop Natural Language Solutions
· Text Analysis: Utilizing Language models for text classification and summarization.
· Speech Solutions: Building conversational AI with Text-to-Speech and Speech-to-Text.
· Translator: Implementing real-time translation agents.

5. Extract Insights from Visual Data (Multimodal AI)
· Vision-Enabled Apps: Developing apps that analyze images and video.
· Multimodal Understanding: Configuring models to produce captions and answer questions based on visual evidence.
· Azure Content Understanding: Configuring pipelines to extract structure from documents and images.
· Document Intelligence: Extracting structured data from forms, documents, and images.
· Azure AI Search: Creating knowledge mining solutions for indexed search.

6. Build Scalable and Production-Ready AI Solutions
· Deployment Strategies: Scaling AI workloads in production environments.
· Monitoring and Logging: Tracking performance, latency, and token usage in Foundry.
· Security: Configuring managed identities and private networking for AI services.