Leadership
The team works on continual learning, inference, product, and operations.




What we believe
“Bring intelligence to the data. The device is the agent. The model is the product.”
Continual learning is our means of putting this understanding into practice.
About
AtomGradient is an on-device AI research company founded by Shuhong Wu. Neural Imprint consolidates local records into learning state that can be saved and restored while base weights remain frozen. We develop inference runtimes, SDKs, and apps. Narrus, Dailyn, and Mealens serve as live validation of our inference runtime; Ururu remains in internal testing.
You can try on-device inference with Narrus, Dailyn, and Mealens, or use our developer tools to integrate your own models and apps.
Offices
Our team works from two locations in China — Beijing and Hefei.
We're looking for researchers and engineers who want to push the boundaries of on-device AI.
View open roles