I am building a next-generation intelligent healthcare platform under the Nevora AI umbrella and need an experienced AI developer to take it from concept to production-ready. Core functionality • Predictive analytics that flags at-risk patients and forecasts outcomes using historical and real-time clinical data. • Natural-language processing that can read, structure and summarise free-text medical notes to enrich decision support. Ecosystem integration The solution must plug seamlessly into existing Electronic Health Records, the telemedicine stack we already run, and a growing catalogue of wearable-device streams (heart-rate, SpO₂, activity). Clean, well-documented APIs are essential so hospital IT can maintain compliance and traceability. User experience Rather than a traditional web dashboard, I want clinicians to interact through an AI medical-assistant “robot” style interface—voice and chat first, with the flexibility to surface critical insights on mobile or desktop when deeper dives are required. Tech expectations Python or TypeScript for core services, TensorFlow/PyTorch for modelling, and a HIPAA-compliant cloud pipeline (AWS HealthLake or equivalent). Please outline your proposed model architectures, data-handling safeguards, and how you will benchmark accuracy, latency and explainability. Deliverables 1. End-to-end AI services (predictive + NLP) ready for containerised deployment 2. Integration modules for EHR, telemedicine API and two wearable SDKs 3. Conversational assistant frontend with authentication and role-based access 4. Test suite, deployment scripts and concise technical documentation Acceptance criteria • AUROC ≥ 0.85 on outcome predictions using withheld validation data • NLP module achieves ≥ 0.90 F1 on entity extraction benchmark • All integrations pass end-to-end data flow tests in a staged environment • Codebase reviewed for security, compliance and handover readiness If you have shipped similar HealthTech products and can demonstrate both AI depth and healthcare interoperability know-how, I’d love to see your approach and a rough milestone plan.