Our vision

Protect your deep learning algorithms in untrusted environments (smartphones, IoT) without sacrificing performance and accuracy.


On-device ML offers the best privacy guarantees as well as an application that will work with no network. But why you should expose your know-how to get it?


Inference can still profit from hardware accelerators and does not suffer accuracy loss

Easy to deploy

You train your model as usual and let the skyld SDK take care of the protection.

SKYLD means…

SHIELD. We make it our top priority to protect your machine learning models and related data. Our Solution acts beforehand to make sure your environment is safe and your model well protected, and after an attack to disable and neutralize the model and related data, to make sure nobody else can ever use it.

SKILLED. Our team gathers high-profile specialists, to ensure our solution is always up to our high standards and to your great expectations. We will join forces with you and be your shield, in these booming yet dangerous artificial intelligence challenges.

Contact us to get a demo!


Our latest blog articles

Is your mobile app vulnerable to extraction?

We can test for free if your mobile application with ML algorithms is vulnearble to extraction attack.

You can contact us at


Want a demo?