Monday, March 21, 2022

Embrace digital transformation at the edge with Azure Percept

Right now, a paradigm shift is underway. Customers are increasingly investing in artificial intelligence (AI) and machine learning (ML) at the edge to transform business operations, deploying intelligent devices that will generate more data in the next three years than in the past 30 years combined.


Of course, transmitting all this data for processing in the public cloud presents challenges like limited bandwidth, latency, storage costs, privacy issues—the list goes on. Only by positioning server-class infrastructure near edge data sources can customers overcome these challenges and generate the real-time insights they need to enable loss prevention, defect detection, traffic management, and countless other scenarios.


Meanwhile, this paradigm shift is blurring the lines between traditional IT and OT organizations. Whether modernizing a conventional datacenter or modernizing a factory, both organizations increasingly find themselves with common needs and priorities.


Yet not all businesses have jumped on the intelligent edge bandwagon. While edge AI is powering digital transformation across every industry, developing solutions can be expensive, complicated, and time-consuming. This is something we’ve heard consistently across device builders, solution builders, operators, and business decision makers.


Which is precisely why we’re investing in a comprehensive lineup that accelerates this journey, whether building Azure AI-ready devices, developing applications running on Azure, or leveraging turnkey SaaS solutions. Below, I’ll share how we’re making it faster and easier to embrace digital transformation at the edge using familiar tools and experiences.


For a deeper exploration, tune into my presentation at NVIDIA GTC, Transforming AI and ML at the edge with Microsoft and NVIDIA.


Simplify AI at the edge with Azure Percept


Fundamental to this transformational shift is the ability to perceive physical environments. That’s where Azure Percept comes in. This fully integrated platform combines three primary components to simplify creating and managing AI edge solutions:


  • Devices: The Azure Percept Edge AI Device Development Kit (DDK) includes the software, tools, and guidance for designing, building, and supporting Azure Edge AI-enabled endpoints. The Azure Percept DDK expedites prototyping, so device builders can quickly assemble, train, and deploy ML modules to tackle multiple outcomes, from detecting workplace hazards to shopper analytics. 
  • Solutions: The Edge AI PaaS provides the building blocks developers need to reduce the time and complexity of creating and operating edge AI applications with solution, device, and security management capabilities across multiple device types.
  • Certification And finally, our Azure Certified Device program makes it easy to discover and differentiate IoT devices built to run on Azure.




Azure Percept on Azure Stack HCI: Built with opportunities in mind


As announced during Ignite, Azure Stack HCI is an integrated virtualization stack that delivers a flexible operating system and the security and performance needed to run virtual machines, containers, and many Azure services on-premises. With Azure Percept on Azure Stack HCI, you can securely deploy, manage, and scale edge AI to build solutions that deliver outcomes.


Furthermore, Azure Percept Studio offers a launchpad for edge models and solutions. Using guided workflows in Azure Percept Studio, you can rapidly integrate AI-capable hardware and Azure AI services to create a proof of concept. Access to both devices and training data expedites prototyping for vision and speech solutions.


Discover what’s new with Azure Percept


Behind the scenes, Microsoft continues to innovate. Here’s a quick roundup of the latest Azure Percept advancements, which I’ll cover in greater detail during my session at NVIDIA GTC.


Azure Percept Edge AI PaaS: Create edge AI applications faster 

Azure Percept Edge AI PaaS streamlines the creation and operation of edge AI applications with solution, device, and security management capabilities. Azure Percept services allow developers to easily deploy and operate AI edge solutions across various device types. Building blocks include:


  • Cloud APIs for building production-ready AI solutions
  • Edge APIs to consume inference results and sensor data and manage insights without direct cloud dependency
  • SDKs to simplify interaction with platform APIs
  • Device identity, authentication, attestation, and security for ML models, end-user data, and other assets
  • Prebuilt user interface elements and code sections

Azure Edge AI Installer: Expedite device setup

To kickstart cloud-ready Azure Edge AI solutions, we’ve deepened integration with NVIDIA Jetson DKs. The Azure Edge AI Installer streamlines device setup for Jetson DKs by automating Azure component installation. The Azure Edge AI Installer will be available in summer 2022 to some Private Preview customers and will support Orin and Xavier Developer Kits.


Azure Edge AI DDK for Device Builders: Build compatible devices

We’re also partnering with NVIDIA and the device builder community to ensure the availability of production-grade devices. The DDK empowers device builders to create secure and compatible devices that accelerate Azure Edge AI solutions from proof of concept all the way through to production.


Azure Certified Device program: Validate and discover compatible devices 

Finally, while not necessarily new, the Azure Certified Device program is unquestionably noteworthy. This program validates and differentiates the capabilities of devices built to run on Azure. The program helps solution builders get started faster and with less investment time while utilizing Azure Percept. Azure Certified Devices make it easier to combine edge products to create business value. Compatible devices can be found on the Azure Certified Device Catalog.


Register for the Azure Edge AI Installer Early Access program


Azure Percept streamlines the secure deployment and management of edge AI resources across IT and OT endpoints. The Azure Percept DDK enables device builders to design and manufacture devices that integrate seamlessly with Azure AI services. The Edge AI PaaS streamlines the creation of secure solutions. And the Azure Certified Device program makes it easy to validate and discover compatible devices with the necessary capabilities.


To get started, sign up for the Azure Edge AI Installer Early Access program at

For more learning and inspiration, I’ve included some bonus resources below.


Azure Percept on Azure Stack HCI 

Azure Percept DK


Posted at