Home > Speakers >

Kwabena Agyeman

Kwabena runs OpenMV, which makes low-power and high-performance embedded machine vision cameras that can run on batteries for years. Over 100K OpenMV Cams are now out in the world. Previously he led all of electrical engineering at Embark Trucks building fully autonomous semi trucks. He has an extensive background in ASICs, FPGAs and microcontrollers.

How AI Accelerated Microcontrollers Will Change Embedded Systems

Status: Coming up in April 2025!

New Microcontrollers available on the market today have megabytes of RAM onboard with AI accelerators providing hundreds of gigaops of computation resources. These new microcontrollers now can replace application processor systems with lower heat generation, power consumption, size, and cost.

The newly announced OpenMV N6 and OpenMV AE3 leverage such Microcontrollers from STMicroelectronics and Alif Semiconductor, which can both run modern YOLO networks onboard at real-time, 30 FPS, frame rates while drawing less than 1W of power.

In the talk we’ll cover how you can use MicroPython to effectively leverage the resources of these new Microcontrollers with Numpy running onboard to efficiently utilize new vector instructions thanks to the new Cortex-M55 processors onboard.

Go to Session