Software company Wind River is collaborating with South Korea’s Nota AI and fellow Californian company SIMa.AI to improve edge AI in real-world deployments.

Nota AI is a pioneer in on-device AI optimisation. The partner programme agreement (PPA) integrates Nota AI’s NetsPresso capabilities into Wind River Studio Developer.

“The combination of technologies from Wind River with its mission-critical expertise for the edge and our proprietary AI model optimisation platform, NetsPresso, will unlock new potential for AI optimisation and on-device generative AI across industries to pave the way for rapid growth and innovation,” said Myungsu Chae, CEO of Nota AI (www.nota.ai). “Together with Wind River, we are accelerating time to market for edge AI applications and significantly expanding the adoption of AI technologies in automotive, mobile, IoT and beyond.”

Avijit Sinha, senior vice president at Wind River (www.windriver.com), added: “AI model optimisation and software-defined automation will be key in bringing to life a wealth of applications and new use cases for the edge. This collaboration, bringing in AI innovation from Nota AI and deep mission-critical and edge expertise from Wind River, will help teams better tap into the power of physical AI and develop the right workflows to explore new possibilities for generative AI on the edge.”

The interoperability of the NetsPresso AI and ML framework with Studio Developer capabilities enables a streamlined workflow for training, compressing, converting, benchmarking and deploying AI models, bridging model optimisation with testing and deployment across real-world edge environments.

NetsPresso (netspresso.ai), Nota AI’s flagship AI model optimisation platform, optimises AI models for device compatibility, allowing them to run directly on the device. Studio Developer (www.windriver.com/studio/developer) is a modern devops platform that accelerates development, deployment, and operation of mission-critical systems for the intelligent edge.

SiMa.ai is a software-centric and embedded edge machine learning system-on-chip (MLSoC) company. It is collaborating with Wind River on integrated hardware and software for edge AI.

Delivering a convenient and seamless experience for developers, SiMa.ai’s MLSoC platform integrated with the enterprise-grade Debian derivative eLxr project with commercial support provided by Wind River’s eLxr Pro, allows developers to customise and accelerate time to production. This integration combines the freedom of open source with enterprise grade security, stability and compliance.

“Edge AI is the next gold rush and creating significant opportunities across robotics, industrial automation, medical, automotive, aerospace and defence,” said Krishna Rangasayee, CEO of SiMa.ai. “Together, with Wind River’s long-standing expertise and market success, SiMa.ai is now jointly delivering the industry’s best edge AI platform that leads in performance, power-efficiency and ease-of-use, addressing all AI needs, including genAI. We are excited to collaborate and jointly enable scaling of AI in all physical industries, on-device.”

Sinha added: “With the complexities that come with the growth of the intelligent edge, AI and software-defined automation will be critical to maximise its potential. The collaboration between SiMa.ai’s advanced MLSoC platform and Wind River’s proven edge technologies in mission-critical environments will help drive real-time physical AI forward across diverse applications and industries.”

The SiMa.ai MLSoC platform is full stack software and silicon, with a unified software development environment called Palette that supports the full ML application development with an effortless create, build and deploy process.

Based on the open source, enterprise-grade Debian derivative eLxr project (elxr.org), Wind River’s eLxr Pro brings commercial enterprise support and maintenance on top of the eLxr community distribution to help users adopt scalable, secure and reliable Linux products that address the complex challenges for cloud-to-edge deployments.