Welcome to Yeti
Our goal at Icii is to drive the vision of AI Everywhere! In particular, we are focused on Source AI, performing the AI processing at the data Source. To do this, we want to make AI as small, low power, and fast as possible. This means pushing AI into smaller and more efficient computing chips. Our current targets are Field-Programmable Gate Arrays (FPGAs).
To run AI models on tiny hardware, chips that don’t even run code, the big question becomes, How? You need to design an architecture that can execute the needed AI operations, schedule the calculations, handle sensor interfaces, and implement model changes. How do you get all of this to work? That’s what Yeti was designed to take care of for you.
Yeti makes running AI models on embedded FPGAs easy. The inputs – your AI model and FPGA. The output – a single file you can program onto your FPGA. That’s it. Nothing complicated. Let Yeti do the heavy lifting, sit back, and enjoy.