I’m building a standalone, AI-powered healthcare platform that combines digital anamnesis with in-clinic workflows. The first milestone is a functional prototype whose core task is diagnosis support: the system collects a patient’s symptom history, runs it through an explainable AI model, and hands clinicians a concise, ranked list of possible diagnoses. Key requirements • Stand-alone architecture – no EMR/HIS integration for now; everything runs in its own Dockerized environment. • Multilingual capability – the NLP pipeline must work in several languages (English plus at least two additional major languages) and be easy to extend. • Dual-facing web UI – patient questionnaire on one side, clinician dashboard on the other, built with a modern JavaScript framework. • Explainability & audit trail – each suggested diagnosis should link back to the data points and model weights that influenced it. • Security & privacy – encrypted data in transit and at rest, HIPAA/GDPR-aligned design patterns, role-based access. • Admin basics – user management, data export, configurable questionnaires. Deliverables 1. Architecture and data-schema document. 2. Trained diagnosis-support model with well-documented inference API (Python + TensorFlow/PyTorch or similar). 3. Working web prototype packaged in Docker (patient & clinician views). 4. Deployment guide/README for local or AWS setup. Feel free to suggest the exact tech stack; I’m comfortable with Python (FastAPI), React/Vue, and standard healthcare data libraries but open to your expert recommendation. When you respond, please include: • A brief overview of similar healthcare or NLP projects you’ve completed. • Your preferred tools/libraries for this build. • A rough timeline broken into major phases. The idea: Patients in the waiting room, via an AI-integrated tablet, fill in their symptoms and receive dynamic follow-up questions. The answers are sent directly to the physician, who can prepare diagnostics and a treatment plan before the consultation. The physical visit will then focus only on: Medical examinations & tests Physical assessment Any remaining questions from the physician Vision: Reduce waiting times by up to 50% by cutting down unnecessary physician time. Create a hybrid model where AI optimizes anamnesis and triage, but the final medical decision-making remains at the physical visit.