Azure AI Foundry is a web portal that unifies various Azure AI services into a single development environment. It integrates the model catalog, prompt flow development, and Azure OpenAI Service capabilities, including model deployment, testing, and custom data integration. Additionally, it connects with Azure AI Services for speech, vision, language, document intelligence, and content safety. [1]
With Azure AI Foundry, teams can efficiently collaborate on AI projects, such as developing generative AI applications using language models. The portal allows for deploying models to real-time endpoints, integrating custom data sources for retrieval-augmented generation (RAG), defining workflows with prompt flow, incorporating content safety filters, and extending solutions with multiple AI capabilities through Azure AI Services. [1]
An Azure AI hub is the foundation for AI projects on Azure, enabling shared assets across multiple projects. Using the AI Foundry and Azure portals, we can create members and assign roles, manage compute instances for experiments, connect to resources like data stores and GitHub, and define policies for managing behavior. [1]
What can I do with a Project?
Projects in Azure AI Foundry are where all AI development happens. We can deploy language models for chatbots, test models in the chat playground, add custom data to prompts, define workflows using prompt flow, evaluate model responses, manage datasets, apply content filters, and use Visual Studio Code in our browser to create and deploy solutions as web apps or containerized services. [1]
It is go-to platform for resource management, team collaboration, and workflow optimization. The platform is particularly suited for developing generative AI applications, offering tools to build content-generating solutions and custom prompt flows. Its extensive model catalog includes options from OpenAI, Microsoft, and Hugging Face, facilitating experimentation with various AI models. Additionally, Azure AI Foundry supports RAG to enhance content relevance and quality. Robust monitoring and evaluation tools ensure models meet performance standards, while seamless integration with other Azure services enables the development of complex, multifaceted projects. Committed to responsible AI usage, Azure AI Foundry provides guidance and tools to adhere to ethical standards and best practices. [1]
Home page looks as follow.
Press enter or click to view image in full size
Press enter or click to view image in full size
Click Customize
Press enter or click to view image in full size
Everything in the following is created newly except Subscription as they are not existed before in my case. Choose location that fits you. Click Next.
Press enter or click to view image in full size
Then, click create.
Press enter or click to view image in full size
Having created Azure AI hub and project, it should look similar to the following image. This is AI hub that I have created and
Press enter or click to view image in full size
Navigate to portal.azure.com and look at resources created within your resource group.
Press enter or click to view image in full size
An Azure AI services, Key vault, Azure AI hub, Azure AI project and stroage account were created under resource group.
Join Medium for free to get updates from this writer.
Subscribe
Subscribe
Now, when we look at Connected resources under Management center, we can select Connected resources, either under our hub or our project, and observe that connections to Azure OpenAI and AI services have already been created.
Press enter or click to view image in full size
Navigate to Models
Press enter or click to view image in full size
Press enter or click to view image in full size
Search for gpt-35-turbo-16k and click confirm.
Give an unique deployment name and click customize. Set the values as below.
Press enter or click to view image in full size
Press enter or click to view image in full size
Click Deploy.
Go to playground in my created project. Click Try the Chat playground.
Press enter or click to view image in full size
Press enter or click to view image in full size
As one can see, many options from chat playground can be selected. In Deployment section, choose the model the we deployed earlier.
In the Chat window, enter What is AI ?
Press enter or click to view image in full size
Azure AI Foundry simplifies AI development by unifying powerful Azure AI tools into a single platform. Whether we’re deploying language models, building workflows, or integrating custom data, the platform offers everything we need to create efficient and scalable AI solutions.
Delete resource group on Azure portal which will consequently delete all the resources created earlier.
Developer SDKs, repos, connectors, and infrastructure helpers. APIs, development tools, and integration resources.
