Microsoft Certified: Designing and Implementing an Azure AI Solution AI-100 Training and Certification

Analyze solution requirements (25-30%)

Recommend Azure Cognitive Services APIs to meet business requirements

• select the processing architecture for a solution
• select the appropriate data processing technologies
• select the appropriate AI models and services
• identify components and technologies required to connect service endpoints
• identify automation requirements/li>

Map security requirements to tools, technologies, and processes

• identify processes and regulations needed to conform with data privacy, protection, and regulatory requirements
• identify which users and groups have access to information and interfaces
• identify appropriate tools for a solution
• identify auditing requirements

Select the software, services, and storage required to support a solution

• identify appropriate services and tools for a solution
• identify integration points with other Microsoft services
• identify storage required to store logging, bot state data, and Azure Cognitive Services output

Design AI solutions (40-45%)

Design solutions that include one or more pipelines

• define an AI application workflow process
• design a strategy for ingesting and egress data
• design the integration point between multiple workflows and pipelines
• design pipelines that use AI apps
• design pipelines that call Azure Machine Learning models
• select an AI solution that meets cost constraints

Design solutions that uses Cognitive Service

• design solutions that use vision, speech, language, knowledge, search, and anomaly detection APIs

Design solutions that implement the Microsoft Bot Framework

• integrate bots and AI solutions
• design bot services that use Language Understanding (LUIS)
• design bots that integrate with channels
• integrate bots with Azure app services and Azure Application Insights

Design the compute infrastructure to support a solution

• identify whether to create a GPU, FPGA, or CPU-based solution
• identify whether to use a cloud-based, on-premises, or hybrid compute infrastructure
• select a compute solution that meets cost constraints

Design for data governance, compliance, integrity, and security

• define how users and applications will authenticate to AI services
• design a content moderation strategy for data usage within an AI solution
• ensure that data adheres to compliance requirements defined by your organization
• ensure appropriate governance of data
• design strategies to ensure that the solution meets data privacy regulations and industry standards

Implement and monitor AI solutions (25-30%)

Implement an AI workflow

• develop AI pipelines
• manage the flow of data through the solution components
• implement data logging processes
• define and construct interfaces for custom AI services
• create solution endpoints
• develop streaming solutions

Integrate AI services and solution components

• configure prerequisite components and input datasets to allow the consumption of Azure Cognitive Services APIs
• configure integration with Azure Cognitive Services
• configure prerequisite components to allow connectivity to the Microsoft Bot Framework
• implement Azure Cognitive Search in a solution

Monitor and evaluate the AI environment

• identify the differences between KPIs, reported metrics, and root causes of the differences
• manage the flow of data through the solution components
• maintain an AI solution for continuous improvement
• monitor AI components for availability
• recommend changes to an AI solution based on performance data

Microsoft AI-102 Exam Syllabus

Select the appropriate Azure AI Service

- Select the appropriate service for a vision solution
- Select the appropriate service for a language analysis solution
- Select the appropriate service for a decision support solution
- Select the appropriate service for a speech solution
- Select the appropriate Applied AI services

Plan and configure security for Azure AI Services

- Manage account keys
- Manage authentication for a resource
- Secure services by using Azure Virtual Ne
tworks - Plan for a solution that meets responsible AI principles

Create and manage an Azure AI service

- Create an Azure AI resource
- Configure diagnostic logging
- Manage costs for Azure AI services
- Monitor an Azure AI resource

Deploy Azure AI services

- Determine a default endpoint for a service
- Create a resource by using the Azure portal
- Integrate Azure AI services into a continuous integration/continuous deployment (CI/CD) pipeline
- Plan a container deployment
- Implement prebuilt containers in a connected environment

Create solutions to detect anomalies and improve content

- Create a solution that uses Anomaly Detector, part of Cognitive Services
- Create a solution that uses Azure Content Moderator, part of Cognitive Services
- Create a solution that uses Personalizer, part of Cognitive Services
- Create a solution that uses Azure Metrics Advisor, part of Azure Applied AI Services
- Create a solution that uses Azure Immersive Reader, part of Azure Applied AI Services

Implement image and video processing solutions (15-20%)

Analyze images

- Select appropriate visual features to meet image processing requirements
- Create an image processing request to include appropriate image analysis features
- Interpret image processing responses

Extract text from images

- Extract text from images or PDFs by using the Computer Vision service
- Convert handwritten text by using the Computer Vision service
- Extract information using prebuilt models in Azure Form Recognizer
- Build and optimize a custom model for Azure Form Recognizer

Implement image classification and object detection by using the Custom Vision service, part of Azure Cognitive Services

- Choose between image classification and object detection models
- Specify model configuration options, including category, version, and compact
- Label images
- Train custom image models, including classifiers and detectors
- Manage training iterations
- Evaluate model metrics
- Publish a trained iteration of a model
- Export a model to run on a specific target
- Implement a Custom Vision model as a Docker container
- Interpret model responses

Process videos

- Process a video by using Azure Video Indexer
- Extract insights from a video or live stream by using Azure Video Indexer
- Implement content moderation by using Azure Video Indexer
- Integrate a custom language model into Azure Video Indexer

Implement Natural Language Processing Solutions (25-30%)

Analyze text

- Retrieve and process key phrases
- Retrieve and process entities
- Retrieve and process sentiment
- Detect the language used in text
- Detect personally identifiable information (PII)

Process speech

- Implement and customize text-to-speech
- Implement and customize speech-to-text
- Improve text-to-speech by using SSML and Custom Neural Voice
- Improve speech-to-text by using phrase lists and Custom Speech
- Implement intent recognition
- Implement keyword recognition

Translate language

- Translate text and documents by using the Translator service
- Implement custom translation, including training, improving, and publishing a custom model
- Translate speech-to-speech by using the Speech service
- Translate speech-to-text by using the Speech service
- Translate to multiple languages simultaneously

Build and manage a language understanding model

- Create intents and add utterances
- Create entities
- Train evaluate, deploy, and test a language understanding model
- Optimize a Language Understanding (LUIS) model
- Integrate multiple language service models by using Orchestrator
- Import and export language understanding models

Create a question answering solution

- Create a question answering project
- Add question-and-answer pairs manually
- Import sources
- Train and test a knowledge base
- Publish a knowledge base
- Create a multi-turn conversation
- Add alternate phrasing
- Add chit-chat to a knowledge base
- Export a knowledge base
- Create a multi-language question answering solution
- Create a multi-domain question answering solution
- Use metadata for question-and-answer pairs

Implement Knowledge Mining Solutions (5-10%)

Implement a Cognitive Search solution

- Provision a Cognitive Search resource
- Create data sources
- Define an index
- Create and run an indexer
- Query an index, including syntax, sorting, filtering, and wildcards
- Manage knowledge store projections, including file, object, and table projections

Implement Conversational AI Solutions (15-20%)

Design and implement conversation flow

- Design conversational logic for a bot
- Choose appropriate activity handlers, dialogs or topics, triggers, and state handling for a bot

Build a conversational bot

- Create a bot from a template
- Create a bot from scratch
- Implement activity handlers, dialogs or topics, and triggers
- Implement channel-specific logic
- Implement Adaptive Cards
- Implement multi-language support in a bot
- Implement multi-step conversations
- Manage state for a bot
- Integrate Cognitive Services into a bot, including question answering, language understanding, and Speech service

Test, publish, and maintain a conversational bot

- Test a bot using the Bot Framework Emulator or the Power Virtual Agents web app
- Test a bot in a channel-specific environment
- Troubleshoot a conversational bot
- Deploy bot logic