I need to automatically generate end-to-end test cases for a WhatsApp-like application whose source code is unavailable. The app supports text messaging, voice calls, and rich media sharing. Although I only have high-level user journeys at the moment, I can describe them in detail when we begin. The solution must run entirely on a local machine—no cloud services—so please lean toward frameworks such as TensorFlow, PyTorch, or any comparable on-premise library you can justify. Your approach should ingest the observable UI flows and user interactions, apply AI/ML techniques to infer test intent, and output executable test scripts that can be replayed through Appium, Selenium, or another open-source driver you recommend. Deliverables: • A repeatable pipeline (code + model) that learns from captured sessions and emits structured test cases. • Setup instructions and sample data so I can retrain or extend the suite on my own. • A brief report explaining the model’s logic, coverage metrics, and how new scenarios can be added. I will consider the job complete once the generated tests reliably reproduce the stated high-level scenarios across the latest Android and iOS builds of the app, and I can run the entire workflow offline on macOS.