Recently, omniml, a start-up company that develops a small-scale fast machine learning model, announced that it has obtained a seed round financing of US $10 million, which will be used to accelerate the deployment of artificial intelligence (AI) on edge devices. The seed round financing was led by GGV capital, and co invested by Qualcomm ventures, footill ventures, matrixpartners, tectonic ventures, GSR ventures, IMO ventures and fellows fund. Omniml plans to use this financing to expand its machine learning team and accelerate software development.

The founding purpose of omniml is to solve the mismatch of computing power between AI applications and edge hardware, so that AI can truly enter all aspects of public life, make home safer and more convenient, make driving safer, and make logistics and factories effectively improve efficiency & hellip& hellip; To achieve these goals, omniml makes the existing machine vision (ML) model smaller and more designed for edge hardware. In this way, these models can directly conduct artificial intelligence reasoning on edge devices, without the need for data centers and cloud environments, greatly reducing operating costs, improving security, ensuring privacy, and being more scalable. Omniml subverts the existing industry, easily empowers the ubiquitous edge artificial intelligence, and improves the speed, accuracy and efficiency of artificial intelligence without additional hardware customization and upgrading

Omniml is jointly led by Dr. Wu Di, co-founder and CEO of the company, Professor Han song, Department of electronic engineering and computer science, Massachusetts Institute of technology, and Dr. Mao Huizi, chief technology officer of omniml. It is worth mentioning that all three of them graduated from the Department of electronic engineering of Tsinghua University. In addition, the developers of omniml technology include the world's top researchers and industry experts from the Massachusetts Institute of technology, Stanford University, University of California, Los Angeles, and engineers who worked for Facebook. It has been proved that it can speed up the main tasks of machine learning by 10 times on various edge devices.