AI-100: Designing and Implementing an Azure AI Solutions

Gain a broad overview of Azure services with hands-on practice via the Azure portal and command line, and learn to describe the most commonly used services in Azure.

Explore Azure compute options—from virtual machines to App Services. Understand storage and databases, including blob storage tiers, Azure Data Lake Storage Gen2, Azure Synapse Analytics, Cosmos DB, and Redis.

Explore the Azure portal to manage resources like virtual networks, virtual machines, and Active Directory, using custom dashboards, services, and a unified console.

Explore the Azure portal to deploy a virtual machine by following a guided wizard—configure subscription, resource group, region, image, size, and networking, and use CLI or PowerShell to manage.

Learn to use the azure cli to provision resources faster than the portal, switch between powershell and bash, and deploy a web app with app services.

Explore Microsoft Azure service categories such as Azure Migrate, Azure Active Directory, DevOps, IoT, analytics, machine learning, Azure Bot Service, and Logic Apps for migration, development, and automation.

Design a scalable Azure solution by integrating App Service, SQL Database, Cognitive Search, AD B2C, queues, Functions, Blob Storage, Redis Cache, CDN, Traffic Manager, and Application Insights.

Learn to monitor, backup, and secure Azure resources using Azure Monitor, Application Insights, and Security Center. Use Azure Advisor, Azure Policy, templates, and Azure Blueprints for governance and cost optimization.

Learn to select a processing architecture for a solution by comparing custom machine learning with pre-built Azure cognitive services, and map data flows and one-way versus two-way conversations.

Organize clean data for cognitive service APIs with Azure tools like SQL Data Warehouse and Data Factory, considering security, size, and format for vision, speech, language, search, and decision.

Select the right Azure AI models and services across vision, speech, search, language, and decision APIs, with examples like computer vision, form recognizer, text analytics, QnA Maker, and content moderator.

Connect cognitive services—vision, search, and decision APIs—into your app or bot via RESTful APIs, SDKs, or CLI, using endpoint and subscription key for secure communication.

Explore Azure automation options for cognitive services, including automation accounts, Logic Apps, and Functions, to deploy workflows and automate actions with endpoints and keys.

Learn how to create an Azure cognitive service in the portal, obtain the endpoint and key, and connect with Visual Studio to run text analysis, facial recognition, and content mediation.

Identify data privacy regulations for AI and cognitive services on azure, including GDPR, HIPAA, PCI DSS, and FedRAMP. Ensure data collection and storage comply with ethical and legal jurisdictions.

Learn how to apply role-based access control for Azure cognitive services using contributor, data reader, and cognitive service user roles, and secure keys with managed identities and Azure Key Vault.

Identify Azure tools to meet security and compliance requirements, including Azure Compliance Manager, Azure Policy, Azure Security Center, and Azure Information Protection, for auditing, reporting, and enforcing regulations PCI, GDPR.

Identify auditing requirements for Azure AI solutions by locating audit trails across Azure AD logs, Azure Activity Logs, diagnostic logs, Application Insights, and Azure Monitor, then verify changes and troubleshoot.

Identify tools for solution, comparing Azure Machine Learning Studio and Azure Machine Learning Service. Leverage Azure Logic Apps and Azure Functions to streamline integration and deployment for ai workflows.

Identify integration points with Microsoft services by orchestrating AI solutions across Azure Functions, IoT Hub, Azure Databricks, Event Grid, Event Hub, and Kafka.

Identify storage for log data, bot state data, and cognitive services outputs using Azure File storage, SQL Database, Cosmos DB, and Azure Monitor.

Train a group-based Azure Face Cognitive Service model to identify and group people in images, create a person group, train, and test with a sample image using keys and endpoints.

Define and implement the ai application workflow: train, package, validate, deploy, and monitor, using azure tools and no-code machine learning studio, with data, connectors, and output integration.

Craft an ingestion and regression strategy using Azure Data Factory, balancing compliance, volume, hybrid deployment, and data transformation for AI and cognitive services.

Design integration points between multiple workflows and pipelines using Logic Apps or Azure Functions with a proxy to trigger, automate, and orchestrate data flows.

Design pipelines by integrating diverse components into a single flow, using cognitive service endpoints and keys to deploy AI models as web services with request, response, and batch request URLs.

Evaluate ai solution cost constraints by analyzing message and transaction rates, compute time, and storage costs, while balancing logic apps triggers and data storage choices.

Design Azure AI solutions using vision, speech, language, knowledge, and search APIs. Compare computer vision with custom vision and apply face, form recognizer, speech, Q&A maker, and anomaly detection capabilities.

Design, test, and deploy bots using bot framework and Azure bots. Learn templates—basic, form, Louis, Q&A, proactive—and cards such as animation and hero, then test with bot emulator and channels.

Discover how language understanding services interpret text or audio inputs into a schema-driven bot output, and learn to design, deploy, and author Lewis bots via Lewis portal or Azure CLI.

Configure bots to connect to multiple channels via the Azure portal and Bot Framework service, enabling web chat and popular integrations like Alexa, Cortana, Facebook Messenger, and Slack.

Integrate bots with azure app services to host the bot service using serverless compute, and configure azure application insights for monitoring, logging, auditing, and troubleshooting.

Explore hardware architectures for AI solutions, comparing CPUs, GPUs, FPGAs, and ASICs on Azure, focusing on cost, parallel processing, flexibility, and machine learning suitability.

Assess when to host machine learning on cloud, on premises, or in a hybrid setup, weighing hardware ownership, Azure services, FPGA options, and data integration with cost and regulatory factors.

Assess and select a compute solution that meets cost constraints by weighing on-premise, cloud, or hybrid architectures, migration options, and control versus regulations, using VMs, containers, serverless, or microservices.

Learn how to authenticate Azure Cognitive Services with subscription keys, tokens, or Azure Active Directory, and set up app registrations, secret, client ID, and API permissions.

Explore how Azure Policy enforces organizational policies for AI designs, offering insight and control to ensure compliance with PCI, DSS, SOC, GDPR, and custom rules.

Design data governance across inputs, outputs, and connected apps using Azure management groups and blueprints to ensure data integrity, compliance, and RBAC across subscriptions and countries.

Design Azure AI solutions with policy, blueprints, and management groups for data privacy and regulatory concerns, saving what's needed and using Application Insights and Log Analytics to minimize retained data.

Learn to extract printed and handwritten text from images and PDFs using the Azure Computer Vision OCR, with multi-language support, via a Python lab setting up API keys and endpoints.

Explore five blocks of an Azure machine learning pipeline—workspace and data store, data reference, compute targets, submit and publish, and run and view results—showing how Azure services fit each stage.

Explore data logging across four intervals in an Azure machine learning project: training, compute creation, image creation, and deployment, with start, stop, and status events for troubleshooting.

Define and construct interfaces for custom ai services by building and training a Custom Vision project, tagging images as domestic or wild, choosing classification or object detection, and testing predictions.

Learn to integrate AI models with other solution components using Logic Apps, Event Grid, API Management, and Service Bus to design scalable Azure AI solutions.

Design solution endpoints that terminate at rest APIs and Azure Event Hub, enabling messages to flow from data producers to consumers like Azure Data Lake Store or Azure Storage.

Learn to consume an api, configure prerequisite components with endpoint and keys, and set up data sets using Azure storage options such as blob storage, file storage, or data lakes.

Explore the differences between KPIs and reported metrics, with planning-phase indicators vs production measurements. Identify roots of discrepancy using Azure Monitor and Application Insights—covering performance, reliability, costs, and bugs.

Identify the differences between expected and actual workflow throughput using Azure Power BI dashboards to compare logs across the train, package, validate, deploy, and monitor stages.

Adopt CI/CD to continuously improve AI solutions by automating testing, building, and deploying code with Azure Pipelines, ensuring seamless updates without production downtime.

Learn how Azure Monitor powers availability for AI infrastructure by integrating web apps, containers, and virtual machines with Application Insights, Log Analytics, and automation for dashboards and reports.

Analyze performance data to recommend changes to an AI solution, scaling up compute and storage, scaling out web apps, and upgrading pricing plans to remove bottlenecks and balance cost.

Review the AI 100 exam skills outline and the Azure AI solutions concepts demonstrated through labs, and provide feedback via email to improve future courses.