The field of AI is developing rapidly and in the last few years impressive research results have been published. Currently available implementations have focused on high end systems, with lots of computing power. We target deeply embedded devices with limited processing power and limited memory. We also offer improvements to more powerful devices where low footprint and minimal performance impact are desired. We make AI possible, where it has not been before.
why are we different
Our software packages bundle several techniques into one integrated system to achieve its goal. At the heart of our software packages is a powerful yet compact neural network inference engine that performs the central AI task. The input and output are processed using other techniques, to reduce the amount of resources necessary to achieve the desired results.
There are no dependencies on other software or libraries. Our software packages are self-contained and can be used directly by your application code without worrying about external dependencies.
Designed with care to work on embedded, without the overhead typical of frameworks running on workstations and servers.
No OS is required, our software can work on a bare-metal environment, RTOS or embedded OS.
Our software packages can be incorporated into your project in a few simple steps. They can be used from C and/or C++ applications.
A deep neural network engine designed specifically for embedded applications. Powerful yet compact. Input and output are additionally processed to improve accuracy, reduce footprint and increase speed.
We offer training classes for your staff, to get up to speed quickly. We tailor our courses to your needs, combining introductory material, theoretical background and product usage.
Do you want to make use of our expertise in your own project? We can help designing and training your DNN, automate your ML and deploy your AI to your embedded target.