Keynote: Modernizing the Development of AI IoT Devices with WebAssembly-based...

on YouTube ↗07m

" summary

Claim: WebAssembly brings memory-isolated, dynamically deployable modules to MCU-class vision sensors. Implementation uses EVP Edge virtualization platform built on WebAssembly Micro Runtime with custom host interfaces for sensors, NN inference (WNN), comms and storage. Development flow provides local CLI plus IDE integration and a Node-RED GUI for quick POCs and offline QA. A purchasable sensor (DSP + MCU) runs a real-time multi-stage NN pipeline: a Japanese license plate reader demo that keeps images on device and successfully produces plate strings.

AI-written summary · about

" related talks

• 07mKeynote: Modernizing the Development of AI IoT Devices with WebAssembly-based...Dan Mihai Dumitriu · Embedded Open Source Summit North America 2024
• 36mWebAssembly-based Edge AI Sensing Pipeline Orches...Kenji Shimizu, Ryu Ishimoto · Emerging OS Forum @ Open Source Summit Japan 2022
• 38mStreamlining AI Solutions Development: S...Shinsuke Tashiro, Munehiro Shimomura · Open Source Summit Europe 2025
• 13mKeynote: WEdge Project: Edge AI enabled by WasmDan Mihai Dumitriu, Tatsuya Kato · WasmCon 2023
• 42mSecure and Efficient Sensing Applications with Wasm: Sony...Ryu Ishimoto, Ekaterina Kravchenko · Open Source Summit Japan + AI_dev 2024
• 20mKeynote: When Wasm Meets Cyber-Physical Systems: A Discussion of WebAssembly in Real-Time, Safety-Critical SystemsEmily Ruppel · WasmCon 2023
• 05mSecure and Efficient Sensing Applications with WasmDan Mihai Dumitriu · WasmCon 2024
• 37mGetting Started with AI and WebAssemblyMichael Yuan, Angel M De Miguel Meana · WasmCon 2023