While building our AI-powered Flutter apps, we always want them to have high fidelity and fast performance. Dart FFI (Foreign Function Interface) enables you to make function calls directly into native code on the device. We can leverage this for faster access local AI and Machine Learn models in our apps.

In this session, we will demonstrate the power of FFI to make Flutter faster with native code for various executions that are fit for those. We will then work through the star on-device model use-case to show how fast it can be. We will profile the app with Flutter DevTools and track performance.

By the end of the session, participants will appreciate the advantages of On-Device AI like privacy, speed and no need for network latency. They will also explore the Dart FFI solution for such direct synchronous access to the On-Device AI models and better understand use cases for these technologies.