AI-102: Designing and Implementing a Microsoft Azure AI Solution – Study Guide

Updated: May 2023 February 2024

There has been an AI specific exam around for a while know, but perhaps the interest in it has been quite limited, not anymore! The AI-102 is becoming very popular, here is what Microsoft have to say about it:

Microsoft Azure AI engineers build, manage, and deploy AI solutions that make the most of Azure Cognitive Services and Azure services. Their responsibilities include participating in all phases of AI solutions development—from requirements definition and design to development, deployment, integration, maintenance, performance tuning, and monitoring.

These professionals work with solution architects to translate their vision and with data scientists, data engineers, IoT specialists, infrastructure administrators, and other software developers to build complete end-to-end AI solutions.

Azure AI engineers have experience developing solutions that use languages such as Python or C# and should be able to use REST-based APIs and software development kits (SDKs) to build secure image processing, video processing, natural language processing (NLP), knowledge mining, and conversational AI solutions on Azure. They should be familiar with all methods of implementing AI solutions. Plus, they understand the components that make up the Azure AI portfolio and the available data storage options. Azure AI engineers also need to understand and be able to apply responsible AI principles.

If you pass the exam, you will earn an associate certification – Microsoft Certified: Azure AI Engineer Associate.

As always, a great place to start is Microsoft Learn. Microsoft now offer their version of a study guide too, this is important to keep an eye on upcoming changes to content, and with AI, I expect this to change more frequently than some other exams. As always, these are free and you can work through them at your own pace. I find this a great way to study and gain greater understanding of the services by actually using them and you will need to be very familiar with Azure networking to pass this exam.

Below I’ve put together a collection of links relevant to the sections Microsoft have highlighted as being part of the skills measured for this exam. These are only guide links, sometimes you need to explore a topic much more deeply if you are not familiar with it. Hopefully these study materials will help guide you to successfully passing AI-102!

Plan and manage an Azure AI solution (15-20%)

Select the appropriate Azure AI service

  • Select the appropriate service for a computer vision solution
  • Select the appropriate service for a natural language processing solution
  • Select the appropriate service for a decision support solution
  • Select the appropriate service for a speech solution
  • Select the appropriate service for a generative AI solution
  • Select the appropriate service for a document intelligence solution
  • Select the appropriate service for a knowledge mining solution
  • https://learn.microsoft.com/en-us/azure/architecture/data-guide/technology-choices/cognitive-services – Cycle through each, familiarise yourself with terms and capabilities of each.

Plan, create and deploy an Azure AI service

Manage, monitor and secure an Azure AI service

Implement decision support solutions (10–15%)

Create decision support solutions for data monitoring and content delivery

Implement computer vision solutions (15–20%)

Analyze images

Implement custom computer vision models by using Azure AI Vision

Analyze videos

Implement natural language processing solutions (30–35%)

Analyze text by using Azure AI Language

Process speech by using Azure AI Speech

Translate language

Implement and manage a language understanding model by using Azure AI Language

Create a question answering solution by using Azure AI Language

Implement knowledge mining and document intelligence solutions (10–15%)

Implement an Azure Cognitive Search solution

Implement an Azure AI Document Intelligence solution

Implement generative AI solutions (10–15%)

Use Azure OpenAI Service to generate content

Optimize generative AI

And that’s it! Good luck with your exam!

Can Azure OpenAI Write This Blog Post?

This is the fear right? Or maybe it’s not a fear, and is a desire for some people, but it’s still not clear what the impact may be. What am I talking about? AI of course; specifically for this post, Azure OpenAI. Now, my idea here is to see what I think of the quality of responses it can give me based on specific prompts to try and tie together a simple blog post. I will pick a topic, an objective for the post, and I will be 100% transparent on what is AI written and what is me – just in case it is not obvious! I will also include the settings being used for my model at the end and whether I adjusted them during.

Ok, let’s get started. My topic is going to be Virtual Network Manager in my usual style of exploration.

What is Azure Virtual Network Manager? Write an opening paragraph for a blog post.

Azure Virtual Network Manager is an essential tool for managing and organizing your cloud-based virtual networks. With virtual networks, you can securely connect and isolate your Azure resources, control traffic flow, and expand your on-premises network into the cloud. However, managing virtual networks can be a challenging task, especially when you have multiple resources and services scattered across different regions. This is where the Azure Virtual Network Manager comes in, offering a centralized and user-friendly interface to help you efficiently manage and monitor your virtual networks. In this blog post, we will take a closer look at the Azure Virtual Network Manager and explore its features and benefits.

What are the top three use cases for Azure Virtual Network Manager? Write a detailed paragraph including bullet points and examples for each.

Azure Virtual Network Manager is a versatile tool that can support various use cases, depending on your organization’s needs. Here are the top three use cases for Azure Virtual Network Manager:

  1. Hybrid Cloud Connectivity: Azure Virtual Network Manager can help organizations establish secure, high-speed connections between their on-premises datacenters and Azure resources. This allows companies to take advantage of the scalability and flexibility of the cloud while maintaining the security and control of their own infrastructure. With Azure Virtual Network Manager, organizations can easily configure VPN gateways, express routes, and other connectivity options to seamlessly integrate their on-premises networks with Azure.

    Examples:
  • Connecting an on-premises Active Directory to Azure to manage identity and access for cloud resources.
  • Extending a corporate LAN to Azure to host applications and services in the cloud.
  • Using Azure as a backup and disaster recovery site for on-premises datacenters.
  1. Multi-Tier Application Architecture: Azure Virtual Network Manager enables organizations to create complex multi-tier application architectures in the cloud, with the ability to isolate and secure different layers of the application. This is done by creating multiple virtual networks, each with its own subnets and security groups. This approach provides better control over network traffic and reduces the risk of security breaches by limiting the attack surface.

    Examples:
  • Hosting a web application in Azure with a separate virtual network for the web server, application server, and database server.
  • Creating a DMZ in Azure to host public-facing services, with a separate virtual network for internal services.
  • Separating development, testing, and production environments into different virtual networks to ensure isolation and security.
  1. Global Network Management: Azure Virtual Network Manager provides a centralized interface to manage virtual networks across different regions and subscriptions. This allows organizations to monitor network traffic, troubleshoot connectivity issues, and optimize network performance from a single location. With Azure Virtual Network Manager, organizations can also implement advanced networking features, such as traffic routing, load balancing, and network security groups.

    Examples:
  • Creating a global load-balancing solution to distribute traffic across multiple regions and availability zones.
  • Implementing Azure Firewall to provide centralized network security management across multiple virtual networks.
  • Enabling Azure Traffic Manager to route traffic to the nearest endpoint based on geographic location or network latency.

What are the current limitations or known issues of Azure Virtual Network Manager? Write a paragraph in the blog post.

While Azure Virtual Network Manager is a powerful tool for managing virtual networks in the cloud, there are some limitations and known issues that organizations should be aware of. Some of the current limitations of Azure Virtual Network Manager include:

  • Limited support for hybrid networking configurations: Currently, Azure Virtual Network Manager has limited support for hybrid networking configurations, which can make it challenging to manage virtual networks that span multiple cloud providers or on-premises datacenters.
  • Limited visibility into network traffic: Azure Virtual Network Manager provides limited visibility into network traffic, which can make it difficult to troubleshoot connectivity issues or monitor network performance.
  • Limited integration with third-party tools: Azure Virtual Network Manager has limited integration with third-party tools, which can make it challenging to implement advanced networking features or security solutions.

    Organizations should carefully evaluate their networking needs and consider these limitations before deploying Azure Virtual Network Manager. However, Microsoft is continually improving the tool and addressing these limitations through updates and new features.

Is there a tutorial or general steps that should be recommended as part of this blog post for people who are interested in now trying Azure Virtual Network Manager? Include high level steps and a link to the appropriate site if possible

Yes, Microsoft offers a comprehensive tutorial on how to set up and manage virtual networks using Azure Virtual Network Manager. Here are the high-level steps for getting started with Azure Virtual Network Manager:

  1. Create a virtual network: The first step is to create a virtual network in Azure. This involves specifying the name, address space, and subnets for your network. You can also configure advanced network settings, such as DNS servers and network security groups.
  2. Create a gateway subnet: If you plan to connect your virtual network to an on-premises network, you need to create a gateway subnet. This subnet is used to deploy a VPN gateway or an ExpressRoute gateway.
  3. Configure VPN or ExpressRoute: Once you have created your virtual network and gateway subnet, you can configure a VPN or ExpressRoute connection to connect your on-premises network to Azure.
  4. Set up network security: Azure Virtual Network Manager provides several tools for securing your virtual network, including network security groups, application security groups, and Azure Firewall. You can use these tools to control inbound and outbound traffic, restrict access to specific resources, and monitor network activity.
  5. Monitor and troubleshoot your network: Azure Virtual Network Manager provides several tools for monitoring and troubleshooting your virtual network, including network performance monitoring, packet capture, and network watcher.

    For more detailed instructions on how to use Azure Virtual Network Manager, check out the official Microsoft tutorial: https://docs.microsoft.com/en-us/azure/virtual-network/tutorial-create-virtual-network.

Based on your previous responses do you think there is anything else that is important to know about Virtual Network Manager? Write a closing paragraph with a call to action to get in touch if you have any questions

In conclusion, Azure Virtual Network Manager is a powerful tool that can help organizations manage and secure their virtual networks in the cloud. With Azure Virtual Network Manager, organizations can establish secure connections between their on-premises and cloud-based resources, create complex multi-tier application architectures, and manage virtual networks across different regions and subscriptions. While there are some limitations and known issues with the tool, Microsoft is continually improving it with updates and new features.

If you have any questions about how to use Azure Virtual Network Manager or want to learn more about how it can support your organization’s networking needs, don’t hesitate to get in touch. Our team of experts is always available to help you navigate the complexities of cloud networking and find the best solutions for your business.

Let’s start with the settings used, I started with these and didn’t change them at all during the process:

engine="35t", (gpt35-turbo)
temperature=0.8,
max_tokens=800,
top_p=0.95,
frequency_penalty=0,
presence_penalty=0

Overall, I think it has the basics nearly correct. However, it has very little personalisation, it obviously speak to no real experience, and is essentially a modified read of the Docs pages. So I am pretty confident that at no point in near future could it be used to write an entire post that I would be happy with publishing. However, where a post calls for something more generic, like known issues, or intros, it could definitely be used as a starting point.

I am interest to see if a GPT4 model would make a huge difference, but for this style of work, I am not sure. Also, I could of course edit the temperature or top p, but I am unsure as to whether that impact again would be noticeable enough.

In conclusion, Azure OpenAI can definitely write a blog post – I am just not sure it’s a blog post I would be happy with!