Allus AI Launches: Vision Foundation Model for Manufacturing

Allus AI recently launched!

‍

^{"Bringing real intelligence to factories."}

‍

^{TL;DR: Traditional manufacturing vision systems are slow, complex, and expensive. Allus delivers a million-dollar vision solution in minutes.}

‍

https://www.youtube.com/watch?v=pNV7Ekyzfm0

‍

Founded by Kai Cui, Shijie Wang & Zhisen (Jason) An‍

Hi everyone, meet Kai, Shijie, and Jason, the cofounders of Allus AI. They’re building a vision foundation model for manufacturing.

‍

The Team

They combine cutting-edge AI with challenges that have troubled manufacturing for over a century. Kai, Shijie, and Jason began exploring how AI could empower industry while studying at Georgia Tech. Kai comes from a family deeply rooted in manufacturing and grew up in factories, managing production for iPhone back panel at the age of 18. Shijie previously worked as an LLM scientist at NVIDIA Lab, where he gained firsthand insight into the pain points of manufacturing. Jason has always operated on the front lines of industrial enterprises and developed universal solutions for robotics companies.

‍

A Vision Foundation Model for Manufacturing

Allus is a 1B parameter vision foundation model trained on over 1.5 billion industrial, robotics, and manufacturing data pairs collected over five years. It achieves state-of-the-art accuracy in defect detection (> 99.95%) and process monitoring (> 99.2%).

Allus understands manufacturing the same way ChatGPT understands language. The dataset they used to train Allus is more than 100x larger than comparable industrial datasets, giving Allus a strong baseline familiarity with the entire factory ecosystem, including machines, people, and processes.

Today 95% of manufacturers still rely on manual inspection to detect defects and monitor process compliance on the factory floor. This work can be automated with computer vision, but these solutions are domain-specific, take months to implement and cost millions of dollars.

With the founders' implementation agent, Allus can be trained to handle specific use cases in only a few minutes using a small number of reference examples.

In the last three months, Allus has been deployed on production lines run by five global fortune 500 manufacturers, where they’ve been able to improve defect detection accuracy in some cases by more than 100x.

‍

Learn More

_‍

_{🌐 Visit allus.ai to learn more.}

_‍

_{🤝 The founders' are working with manufacturers across the globe, from electronics and automotive to food & beverage, FMCG, and high-performance materials.}

‍

_{📧 If you want to bring intelligence to your factory, talk to them! Reach out via email here.}

_‍

_{👣 Follow Allus AI on LinkedIn & X.}‍
‍