Cross-Platform ML: Native SDKs, TensorFlow, and KMP in Action

This session shares our on-going experience implementing on-device TensorFlow ML capabilities through Kotlin Multiplatform to serve both our Android and iOS native SDKs. We'll explore the technical challenges around KMP and the current state of ML on KMP, architectural decisions around common KMP integration issues like code management, iOS framework binding and outputs. Learn how we brought a team of data scientists into a Kotlin-based workflow and leveraged AI tooling to accelerate their learning curve, established efficient code practices, and relied on comprehensive testing for ML code using KMP.