Artificial intelligence at the edge: Building in-demand IoT skills
Build in-demand IoT skills by understanding artificial intelligence at the edge. Here's why it matters for developers and more about AI and IoT in Azure.
Jun 08, 2023 • 10 Minute Read
In this blog post, we’ll talk about AI at the edge and why it matters to developers; the core technologies to support Applied AI and IoT within the Azure ecosystem; and how to get started learning about Microsoft Azure and IoT and skilling up on Applied AI for IoT.
IoT skills are in high demand
The global market for IoT (Internet of things) is growing exponentially, from $250B in 2020, to a projected 1.6T by 2025. The use of AI (artificial intelligence) with IoT is one of the identified growth trends with a rising need for three types of skills:
- Processing large volumes of IoT data (MLOps, Analytics)
- Integrating cloud-based intelligence at scale (Cognitive Services, IoT Hub)
- Deploying more AI-enabled IoT devices (IoT Edge)
What does this mean to you as a developer? It means there is demand for IoT engineers with AI experience and knowledge — and now’s the perfect time to skill up.
In this blog post, we’ll focus on the topic of Artificial Intelligence at the Edge and leave you with some great resources to jumpstart your IoT exploration, including official course content from Microsoft Learn and A Cloud Guru resources that can help you earn your AZ-220 Azure IoT Developer Certification.
This post is meant to help:
- Create awareness of relevant resources for everyone
- Jumpstart learning journeys into IoT for IoT beginners
- Skill up on AI technologies for IoT for IoT practitioners
Let's dig in!
Become an Azure IoT Developer
Get hands-on to learn about Azure IoT Solution Infrastructure and prepare for the Microsoft Azure IoT Developer (AZ-220) exam with ACG.
What is artificial intelligence at the edge?
Artificial intelligence at the edge is an emerging field that combines the latest breakthroughs in AI and machine learning to drive hardware solutions that often require a need for mission-critical insights in near real-time without the need for full connectivity to the internet.
These self-contained solutions are often powered by hardware-accelerated devices but can also be employed on tiny microcontrollers.
This subject of AI at the edge was one of the themes for #JulyOT, 31 days of content and projects to inspire those curious about IoT to pursue personal projects within the IoT realm and share their learnings with the community to motivate others. (Check out the #JulyOT post at the Microsoft IoT Tech Community to learn more.)
A visual guide to Azure Percept
Azure Percept is an emerging family of hardware, software, and services from Microsoft, designed to accelerate innovation and deployment of business solutions using IoT and AI at the edge.
The visual guide below provides a summary of the core components of Azure Percept (Dev Kit, Studio, and Best Practices) and emphasizes its silicon-to-services approach to exploring IoT scenarios beyond telemetry.
Microsoft’s #JulyOT content covers a wide range of these scenarios with a particular focus on the Azure Percept DevKit from Microsoft, enhancing Machine Learning models using Transfer Learning strategies powered by NVIDIA hardware, and an AI-driven nose developed using a Microcontroller and Machine Learning solutioning from Edge Impulse.
Explore the full series of content around “Artificial Intelligence at the Edge” — and go from learning core applied AI concepts, to exploring fully functional projects with code using Azure services.
IoT and AI at the edge learning resources
Retro Game Translation Tablet with C#, Cognitive Services, and Azure IoT Edge | Learn how Azure IoT Edge enables you to deploy cloud intelligence locally on edge devices with a demo showcasing Cognitive Services Containers and retro videogames! |
Azure Percept DevKit Getting Started | Interested in Applied AI for IoT? Learn about Azure Percept Dev Kit, and use Azure Percept Vision and Audio modules in your solutions! |
First Impressions of the Azure Percept DK | Let’s unbox the Azure Percept Dev Kit (DK), learn about its components and setup — and get started training your first AI model! |
Perceptmobile: Azure Percept Obstacle Avoidance LEGO Car | Azure Percept simplifies use of Azure AI technologies on the IoT edge. Learn how you can build a Percept-enabled obstacle-avoidance solution for LEGO Boost cars (with code!) |
Azure Percept Audio - First Steps | Dive into the Azure Percept Audio (or Percept Ear) module, enabling audio processing at the edge. Explore built-in demos and templates! |
Train smarter with NVIDIA pre-trained model and Transfer learning Toolkit on Microsoft Azure | Edge devices are resource-challenged. Learn how NVIDIA’s Transfer Learning Toolkit (TLT) accelerates development & builds efficiency. |
Edge Impulse "One More Epoch" Livestream with Benjamin Cabe | Edge Impulse is a development platform for ML on edge devices. Learn how to use it to build an artificial nose that senses smells in real-time! |
30 Days to Learn It: Beginner-Friendly!
Now that you have a sense for the value of applied AI for IoT, let's talk about resources to get you started with your own IoT journey as an absolute beginner.
We want to challenge you to sharpen your knowledge of Azure IoT Services with the Azure IoT Developer Journey — designed to guide learners in pursuit of an official designation as a certified Azure IoT Developer.
Need further incentive? Try the “30 Days to Learn It Cloud Skills Challenge,” a limited-time promotion that challenges you to learn, and apply, your Azure IoT knowledge by completing a curated series of interactive learning modules from the Microsoft Learn Online Learning Platform.
Register for this challenge and — if you complete the assigned modules within 30 days — you may be eligible to receive a 50% discount voucher to take the official AZ-220 IoT Developer Certification Exam.
Here’s what you’ll learn in this Azure IoT Developer Collection (18 modules, 15 hours) of content:
1. Enabling Digital Transformation - understand what digital transformation means and how it helps drive business solutions using Cloud, AI and IoT. |
2. Microsoft Azure IoT Strategy and Solutions - understand the broader Azure IoT Landscape - the tools and services it provides and how to use it to build and deploy IoT-powered solutions. |
3. Introduction to Azure IoT Hub - understand how to deploy and manage large-scale IoT solutions using Azure IoT Hub and component services. |
4. Learn how to manage IoT devices as an IT Admin - understand how to configure Azure IoT Hubs, register and run IoT devices – and administer IoT deployments for your organization. |
5. Create your first Azure IoT Central App - Get hands-on learning by building an Azure IoT Central App to monitor and command a refrigerated truck – learn to navigate the Azure IoT portal! |
6. Introduction to Azure IoT Edge - learn about IoT Edge, its capabilities and components, and the problems it solves. Understand when and how to use this in your IoT solutions. |
7. Introduction to implementing lambda arch. solutions for IoT - learn about hybrid lambda IoT architectures — explore Azure storage service options (Blob, Data Lake, Cosmos DB) and analytics (Time Series Insights) services and learn when, and how, to use them. |
8. Explore & Analyze Time-Stamped Data With Time Series Insights - understand how to collect, process, store, analyze, and query – your IoT data at scale, using insights from this service. |
9. Automatically provision IoT devices securely, and at scale, with Device Provisioning Service - create a Device Provisioning Service (DPS) to securely handle multiple remote devices. |
10. Identify anomalies by routing data via IoT Hub to a built-in ML model in Azure Stream Analytics - Create an app to simulate conveyor belt vibration. Use it to explore routing, anomaly detection. |
11. Remotely monitor and control devices with Azure IoT Hub - Use “monitoring & controlling temperature/humidity in a cheese cave” as a scenario to create IoT app (device) + service (cloud). |
12. Automate IoT devices management with Azure IoT Hub - learn to use IoT Hub to simplify complex IoT device management processes (e.g., firmware updates for device groups) etc. |
13. Manage your Azure IoT Hub with alerts and metrics - understand metrics, alerts, diagnostics and logs. Use vibration telemetry as a use case to create an IoT Hub app and test alerts & metrics. |
14. Deploy a pre-built module to the Edge device - use pre-built temperature simulator module as example, deploying it to IoT edge device (in container) and viewing simulated data (in portal). |
15. Set up an IoT Edge Gateway - understand what functionality an IoT gateway provides and walk through how an IoT Edge device can be configured for use as a gateway. |
16. Set up rules and take action on telemetry data in Azure IoT Central - builds on the refrigerated truck example (in #5), creates rules and actions to respond to telemetry inputs for automation. |
17. Manage IoT Central applications with the REST API - describes how to manage your IoT apps programmatically – to add device templates, create devices, set properties, send commands... |
18. Set up continuous data export from Azure IoT Central to a Power BI app - builds on refrigerated truck example (#5, #16) to add continuous data export to Azure Blob Storage & event hub. |
AI at the edge: Specialization path
The above learning path offers a curated set of modules to jumpstart your IoT learning journey with Azure-powered technologies. If you already have an understanding of the IoT ecosystem, then explore the modules in this section to build your knowledge and experience with applied AI at the edge.
The content has been created by Microsoft in a working partnership with the University of Oxford and includes modules leveraged in the university's official Artificial Intelligence: Cloud and Edge Implementations course. Learn a complex subject using materials from one of the best programs on the planet — and take your interest in AI at the Edge to the next level.
Check out this collection to get started — click on this link to get the interactive version of the roadmap below with a higher level of detail.
Want a quickstart? Explore the intro modules identified below, and the two options for integrating Cognitive Services with IoT solutions – on the backend (triggered by Azure Functions, in the cloud) or on the IoT edge device (created on cloud, deployed to Edge).
AI Edge Engineer: Learning Collection | 17 modules, 17.5 hours |
Introduction to Azure IoT | Understand what problems Azure IoT solves, and the components and services it provides for devs. |
Introduction To Azure IoT Hub | Understand the components and capabilities of Azure IoT Hub and learn how (and when) to use it. |
Introduction to Azure IoT Edge | Understand what IoT Edge provides (the modules, cloud interface, runtime) and when to use it. |
Introduction to Azure Functions For IoT | Understand how to leverage triggers and bindings to create scalable and serverless IoT solutions. |
Introduction to MLOps for IoT Edge | Understand how MLOps automates development & deployment pipelines for ML models on IoT edge. |
Introduction to Azure Sphere | Azure Sphere is a platform to develop, deploy & maintain secure internet-connected IoT solutions. |
Connect IoT Devices To Cognitive Services Using Azure Functions | Running machine learning services in the cloud, triggered from IoT device as serverless functions. |
Run Cognitive Services on IoT Edge | Create the cognitive service in the cloud, configure and containerize it — deploy to IoT edge device. |
Summary and next steps
The IoT ecosystem is growing rapidly, and developer skills in IoT and AI/ML will be critical to supporting next-generation IoT deployments and addressing applied AI challenges in data processing, cognitive service integrations, and intelligence at the edge.
Want to skill yourself up to be ready for this challenge? Here are some actions you can take:
- Check out Microsoft's #JulyOT content page to build your awareness of IoT terms and technologies.
- Try the #30DaysToLearnIt challenge to go from IoT concepts to code — and aim for certification!
- Complete Microsoft's AI Edge Engineer learning path and skill up on AI integration with IoT cloud and edge!
Keep an eye out for the next blog post from our team that focuses specifically on “Beginners, Students, Teachers and Makers” anchored by an amazing 24-lesson curriculum with hands-on projects (and an associated kit/bundle from Seeed Studio) for the quickstart.
Until next time, happy hacking!
About the author
Paul DeCarlo is a Microsoft Sr. Cloud Developer Advocate focusing on IoT, OSS, Containers, and Kubernetes. Nitya Narasimhan also contributed to this post.
Azure your success in the cloud
Get certified, master modern tech skills, and level up your cloud career — whether you’re starting out or a seasoned pro. Learn by doing with ACG.