Implement Azure Edge AI Accelerator
Budget: $10 – $30 USD
I need a practical, demonstration-ready solution that proves how an Edge AI accelerator can be harnessed inside Azure. The main objective is to keep inference on the device using Azure Edge AI capabilities while orchestrating everything through Azure IoT Hub for registration, telemetry and over-the-air updates. Example is already given in this Repo - https://github.com/Azure-Samples/edge-aio-in-a-box
Scope
You will package a representative model, optimise it for the accelerator and deploy it to an edge device via IoT Hub. The exact model domain—computer vision, NLP or predictive analytics—is flexible; the priority is to show low-latency, on-device inference with a clean end-to-end pipeline.
Example is already given in this Repo - https://github.com/Azure-Samples/edge-aio-in-a-box
Deliverables
• Source code for the edge application and model wrapper use case and any simulated data for sensor data
• Deployment artefacts (Dockerfile, IoT Edge deployment manifest, IaC scripts)
• A concise walkthrough that lets me reproduce the setup, push updates and verify performance
• need to make the use case to be tunning in my machine in my Azure Subscription with out any issues.
The build is complete when the model runs locally on the hardware, metrics flow into IoT Hub, and I can swap in an updated model without physical access to the device.
Scope
You will package a representative model, optimise it for the accelerator and deploy it to an edge device via IoT Hub. The exact model domain—computer vision, NLP or predictive analytics—is flexible; the priority is to show low-latency, on-device inference with a clean end-to-end pipeline.
Example is already given in this Repo - https://github.com/Azure-Samples/edge-aio-in-a-box
Deliverables
• Source code for the edge application and model wrapper use case and any simulated data for sensor data
• Deployment artefacts (Dockerfile, IoT Edge deployment manifest, IaC scripts)
• A concise walkthrough that lets me reproduce the setup, push updates and verify performance
• need to make the use case to be tunning in my machine in my Azure Subscription with out any issues.
The build is complete when the model runs locally on the hardware, metrics flow into IoT Hub, and I can swap in an updated model without physical access to the device.