I want to bring a studio-grade, AI-powered music separation app to Android. The concept is straightforward: the user selects an MP3 or MP4, taps a button, and the app returns isolated stems—vocals, drums, bass, piano, and a clean instrumental track—ready for practice, remixing, or karaoke. Key points I need you to cover • A simple, minimalistic UI: one-screen flow for selecting the file, optional audio preview, and a clear progress indicator during separation. • File handling restricted to MP3 and MP4 on the first release. • Fast, high-quality stem extraction. Whether you rely on TensorFlow Lite, PyTorch Mobile, or another on-device solution is up to you, but the end result must rival current web-based services. • An in-app audio preview so the user can listen to the original and the separated stems before saving. • Clean architecture, Kotlin-first (Java acceptable for libraries), and clear documentation so future maintenance is painless. Deliverables – Signed APK and Play-ready bundle – Complete, well-commented source code with build instructions – Integrated or referenced AI model files – A short test report demonstrating successful separation on at least three sample MP3/MP4 tracks If you have previous work with audio DSP or mobile ML, please mention it when you respond. I’m ready to move quickly and can provide sample tracks the moment we agree on the approach.