OmniML Launches Platform that Reimagines MLOps for Edge AI

Omnimizer greatly reduces the time and cost of deploying AI to the edge by adapting models to fit diverse hardware platforms

SAN JOSE, Calif.–(BUSINESS WIRE)–OmniML, a startup making artificial intelligence (AI) more accessible to all edge devices, today announced the release of Omnimizer™, a platform that simplifies and accelerates machine learning operations (MLOps) by bridging the gap between ML models and edge hardware.

AI is becoming ubiquitous, powering everything from self-driving cars to nearly every application on mobile devices. However, many ML models are not well-suited for edge devices because they were born in the cloud and do not fit those devices well. Such a mismatch greatly hinders the potential of AI and has been a universal barrier for companies at the frontier of the edge AI revolution.

Omnimizer solves this mismatch by auto-adapting and optimizing models for hardware, allowing ML engineers to focus on model design and training without worrying about the complex details of hardware deployment. It gives deployment engineers the confidence that the models will work properly on target devices without the need for redesign and iterations between multiple teams. As a result, Omnimizer eliminates the inefficiencies that lead to slow deployment, poor performance, and higher costs.

A maker of smart cameras, for example, used Omnimizer to significantly reduce the complexity of designing a model for its low-cost chips, reducing deployment time drastically while achieving superb inference performance on its edge devices.

“There is so much promise in using AI closer to where people live, but it is still too inefficient and costly to deliver these tremendous benefits for everyone,” said OmniML Co-Founder and CEO Di Wu, PhD. “Omnimizer solves this by unifying workstreams of ML development and deployment, enabling enterprises to adapt existing models for their hardware based on their specific business needs.”

The magic behind Omnimizer’s capability comes from the inventors of “deep compression” and its all-star founding team that pioneered industry-leading technologies in hardware-aware neural architecture search (NAS).

But this technical strength is just the beginning. In the year since the company’s inception, the OmniML team has been perfecting Omnimizer and serving industry-leading companies in electrical vehicles, autonomous driving, robotics, smart cameras, and many more. Many ML productivity features were added to empower users to profile, diagnose, optimize, and prototype ML models for edge hardware deployments effortlessly.

Supporting most machine learning capabilities for Computer Vision, Natural Language Processing, and many other domains, Omnimizer can intake open-source or a customer’s existing ML models with a few lines of code. It is powered by a cloud-native backend infrastructure that enables hassle-free model adaption and deployment for almost all major chip platforms including CPUs, GPUs, and AI SoCs.

“Our work with OmniML reflects our continued commitment to develop powerful, next-generation software, advanced machine learning and seamless AI solutions to further advance robotics through platforms such as the Qualcomm® Robotics RB5 Platform,” said Dev Singh, senior director of business development and head of autonomous robotics, drones & intelligent machines at Qualcomm Technologies, Inc.

Current Omnimizer customers are leveraging the platform to optimize ML models for autonomous vehicles, robotics, IoT, and mobile devices. OmniML is also working on proof-of-concept opportunities in industrial automation, smart appliances, and pharmaceuticals, among other industries, as part of its mission to bring the benefits of AI to everyone.

About OmniML

OmniML’s AI software enables enterprises that work on mobile, automotive, robotics, and IoT applications to produce small, fast, and energy-efficient ML models for on-device AI. OmniML is backed by established VCs including GGV Capital, Qualcomm Ventures, and others. It is led and founded by world-leading researchers and industry experts including Dr. Song Han, MIT EECS Professor and experienced start-up founder, Dr. Di Wu, former Facebook engineer, and Dr. Huizi Mao, co-inventor of the “deep compression” technology that came out of Stanford University. OmniML solves the mismatch between AI development and hardware deployments to make AI more accessible everywhere. Founded in 2021, OmniML is headquartered in San Jose, California.


Media Contact:
Henry Guo

error: Content is protected !!