I am creating a text-based AI medical assistant that lets patients chat directly with a doctor online. The core of the project is the conversational engine: it must understand user symptoms, triage appropriately, and seamlessly hand the conversation to a live physician when needed. All interactions stay in text for now, but I still want the dialogue to feel natural and reassuring. Please train or fine-tune the model so it answers safely within medical best-practice boundaries, flags emergencies, and clearly states when a human professional is taking over. Language support is limited to the major markets we are targeting, so every feature you build has to work fluently in both English and Chinese. That includes intent detection, follow-up questioning, and any templated doctor responses. Deliverables need to cover: • A deployable chatbot package with the bilingual NLP pipeline in place • Simple web chat interface or embeddable widget for integration into my existing site • Clear hand-off logic and sample doctor dashboard/messages so the transition from AI to human is smooth • Brief technical documentation explaining model choices, training data requirements, and how to extend to more languages later I will test by running common symptom scenarios in both languages, checking accuracy, clarity, and the hand-off process. Once the AI passes those scenarios without hallucinating or missing red-flag symptoms, the milestone is complete.