Exam AI-103: Developing AI Apps and Agents on Azure (Video)

Description

Master the art and engineering of building production-grade AI applications and autonomous agents on Azure.

The AI-103 is the successor to the beloved AI-102 (Azure AI Engineer Associate), and it represents a quantum leap in scope and ambition. Where the AI-102 focused on consuming individual Azure AI Services--a bit of vision here, some NLP there--the AI-103 is built around Microsoft Foundry as the unified platform for building, deploying, and operating AI applications and autonomous agents. The exam demands that you plan Azure infrastructure for AI, implement RAG pipelines, build multi-agent orchestrations with safeguards, and operationalize the whole stack with tracing, token analytics, and safety monitoring.

This is a developer-focused course, and the approach reflects that: approximately 20% conceptual architecture slides and 80% live-coded, screencast-based demonstrations. Every lesson follows a consistent pattern: explain the architectural context, build the solution live in Foundry and VS Code, then review what we built and connect it to exam objectives.

Skill Level:

• Intermediate to Advanced

Learn How To:

• Choose the right model for each task--LLMs, small language models, multimodal models, and Foundry Tools
• Design Azure infrastructure for AI apps and agent-based solutions with proper security and CI/CD integration
• Configure safety filters, guardrails, risk detection, and content moderation for responsible AI
• Deploy and consume models via Foundry, implementing RAG with semantic, hybrid, and vector search
• Build agents with retrieval, function-calling, conversation memory, and multi-agent orchestration
• Implement model reflection, chain-of-thought evaluations, and self-critique loops
• Build image and video generation solutions with editing workflows including inpainting and mask-based edits
• Implement speech-to-agent interactions with custom speech models and multimodal audio reasoning
• Build retrieval and grounding pipelines with content ingestion, indexing, and enrichment
• Extract structured information from documents using multimodal pipelines combining OCR, layout analysis, and field extraction

Course requirement:

Pre-requisites:

• Intermediate Python development skills (functions, classes, async patterns)
• Experience consuming REST APIs and working with SDKs
• Working knowledge of Azure services (resource groups, storage accounts, networking basics)
• Familiarity with AI and ML concepts (models, inference, training data)
• AI-901 certification recommended but not required
• An Azure subscription with pay-as-you-go billing enabled--demos use pay-per-call model deployments, agent runs, and media generation. Apply a budget across the course, and run the cleanup script at the end of each lesson.

Who Should Take This Course:

Job titles:

• Azure AI Engineers building and deploying AI solutions
• Full-stack developers integrating AI into applications
• Cloud architects designing AI infrastructure
• DevOps engineers operationalizing AI workloads
• Data engineers building RAG and retrieval pipelines

About Pearson Video Training:

Pearson publishes expert-led video tutorials covering a wide selection of technology topics designed to teach you the skills you need to succeed. These professional and personal technology videos feature world-leading author instructors published by your trusted technology brands: Addison-Wesley, Cisco Press, Pearson IT Certification, Sams, and Que. Topics include IT Certification, Network Security, Cisco Technology, Programming, Web Development, Mobile Development, and more. Learn more about Pearson Video training at http://www.informit.com/video.

Sample Content

Table of Contents

Choose Foundry Services, Models, and Integration Approach

    Choose an appropriate model for each task, including LLMs, small language models, multimodal models, and Foundry Tools

    Choose the appropriate Foundry services for generative tasks, grounding, vector search, agent workflows, or multimodal processing

    Choose an appropriate method for retrieval and indexing

    Choose appropriate memory, tool, and knowledge integration services for agent solutions

Set Up AI Solutions in Foundry

    Design Azure infrastructure for AI apps and agent-based solutions

    Choose appropriate deployment options

    Configure model and agent deployments

    Integrate Foundry projects with CI/CD pipelines

Manage and Monitor AI Systems

    Manage quotas, scaling, rate limits, and cost footprints for model and agent workloads

    Monitor model performance, drift, safety events, and grounding quality

    Monitor data ingestion quality, search index health, and relevance performance

Secure AI Systems with Zero Trust Identity and Networking

    Configure security, including managed identity, private networking, keyless credentials, and role policies

    Apply Zero Trust networking principles to Microsoft Foundry projects by using private endpoints and virtual network integration

    Assign built-in and custom Azure RBAC roles to service principals, users, and managed identities accessing Foundry resources

Implement Responsible AI Across Generative and Agentic Systems

    Configure safety filters, guardrails, risk detection, and content moderation

    Apply responsible AI instrumentation, including evaluators, safety evaluations, and explanation tooling

    Implement auditing through trace logging, provenance metadata, and approval workflows

    Govern agent behavior with oversight modes, constraints, and tool-access controls

Build Generative Applications with the Foundry SDK

    Deploy and consume LLMs, small models, code models, and multimodal models

    Integrate generative workflows into applications by using Foundry SDKs and connectors

    Configure an application to connect to a Foundry project

Prompt Engineering and Optimizing Generation Behavior

    Tune generation behavior, such as prompt engineering and adjusting model parameters

    Apply prompt engineering patterns including role prompting, few-shot examples, structured output formatting, and grounding instructions

    Implement model reflection, chain-of-thought evaluations, and self-critique loops

Implement Retrieval-Augmented Generation and Multistep Reasoning Workflows

    Implement RAG in an application

    Design workflows, tool-augmented flows, and multistep reasoning pipelines

    Orchestrate multiple models, flows, or hybrid LLM and rules engines

Evaluate, Observe, and Operationalize Generative AI Apps

    Evaluate models and apps, including detecting fabrications, relevance, quality, and safety

    Set up observability by implementing tracing, token analytics, safety signals, and latency breakdowns

    Instrument Microsoft Foundry applications with Azure Monitor, Application Insights, and OpenTelemetry for end-to-end production telemetry

Build Agents with Foundry: Roles, Memory, Tools, and Function

    Define agent roles, goals, conversation-tracking approach, and tool schemas

    Build agents that integrate retrieval, function-calling, and conversation memory

    Integrate agent tools, including APIs, knowledge stores, search, content understanding, and custom functions

Orchestrate Multi-Agent Solutions and Monitor Agent Behavior

    Implement orchestrated multi-agent solutions

    Build autonomous or semiautonomous workflows with safeguards and approval flow controls

    Integrate monitoring into deployed agents, evaluate agent behavior, and perform error analysis

Generate and Edit Images and Videos with Foundry

    Implement a solution that generates images from text prompts and reference media

    Implement a solution that generates videos from text prompts and reference media

    Configure image-editing workflows, including inpainting, mask-based edits, and prompt-driven modifications

    Implement workflows to edit generated videos

    Select and apply appropriate generation and editing controls provided by the platform

Multimodal Understanding and Visual Analysis Workflows

    Build a solution that analyzes visual context by using multimodal models

    Configure apps to produce concise or detailed captions for single or multiple images

    Implement a solution that enables question-answering grounded in visual evidence

    Configure generation of alt-text and extended image descriptions aligned to accessibility guidelines

    Implement visual understanding by configuring Azure Content Understanding in Foundry Tools to extract visual characteristics

    Implement video analysis workflows to process and interpret video segments

    Configure single-task and pro-mode content-understanding pipelines

    Implement solutions that identify objects, components, or regions within images or video

Implement Responsible AI for Multimodal Content

    Implement filters to classify unsafe or disallowed visual content

    Detect and mitigate indirect prompt injection by using embedded text in images

    Enforce visual policy rules, such as applying watermarks, flagging prohibited symbols, upholding brand usage requirements, and detecting inappropriate content

Apply Language Model Text Analysis and Translation

    Implement solutions to extract entities, topics, summaries, and structured JSON outputs by using generative prompting and Foundry Tools

    Configure detection of sentiment, tone, safety issues, and sensitive content

    Build solutions that translate text by using Azure Translator in Foundry Tools or LLM-powered translation flows

    Customize language model outputs for domain tasks, such as compliance summarization and domain extraction

Implement Speech Solutions for Agentic Interactions

    Implement workflows to convert speech to text and text to speech for agentic interactions

    Integrate speech as an agent modality, including custom speech models

    Enable multimodal reasoning from audio inputs

    Translate speech into other languages by using language models and Foundry Tools

Build Retrieval and Grounding Pipelines

    Ingest and index content, such as documents, images, audio, and video

    Configure semantic search, hybrid search, and vector search for grounding

    Implement enrichment by using custom or built-in skills for text, images, and layout

    Configure RAG ingestion flow, including documents and using OCR

    Connect retrieval pipelines directly to workflows and agent tools

Extract Content from Documents with Content Understanding

    Extract information by using multimodal pipelines that combine OCR, layout analysis, and field extraction

    Produce clean, grounded representations to use with agents and RAG by using Content Understanding

    Implement analyzers for generating structured or markdown outputs for downstream reasoning by using Content Understanding

SUMMARY

Updates

Submit Errata