Business Advantages

KanduAI Technology

​

To make our solutions powerful, practical and scalable (low cost, small form factor) we have developed unique technology, AI Inference Engine and SDK for low power CPU.
Our patent pending method (United States Patent Application No. 16/232,685) reduces the amount of calculations required to achieve state of the art accuracy and speed.
 

On top of our patent pending ML inference engine for embedded devices we have built multiple state of the art models required for the real world use cases. These models include Human detection,Face detection ,Face recognition ,Gender classification,Age estimation, Generic Object detection & classification and are now being used by our clients for various use cases.

A Groundbreaking Convolutional Neural Network Architecture

KanduAI's technology is designed for computational, memory and power efficiency achieving state-of-the-art precision while using less than 4% of resources used by current advanced models.

The architecture and reduction in resources usage minimizes processor and memory bandwidth requirements and allows cutting edge AI functionality on low-end IoT devices.