AI Chatbot for Elder Care

We are seeking an experienced AI/LLM Engineer to design and implement a secure AI chatbot that supports caregivers within our Elder Care Management application. Since our application processes sensitive health-related patient information, the solution must be designed with data privacy and security by design and comply with applicable Swiss data-protection requirements. A key architectural requirement is that sensitive patient data must remain within Switzerland. The objective is to allow authorized caregivers to interact with resident information using natural language. Example use cases could include: "What happened with Mrs. X during the last 7 days?" "Show me the latest care reports for Mr. Y." "Has Mrs. X had any falls recently?" "What are the current care goals for this resident?" "Summarize the relevant information for the shift handover." The AI should retrieve and summarize information from the patient's existing care documentation while considering the user's access rights. We are particularly interested in candidates who can propose a technically sound architecture that provides modern LLM capabilities without exposing sensitive patient information to AI services outside the permitted environment. Scope of Work The selected candidate would initially be expected to: - Review our existing application architecture and data model. - Propose the AI/RAG architecture for securely accessing resident information. - Recommend suitable LLM and infrastructure options considering the Swiss data-residency requirement. - Implement a proof of concept. - Implement secure retrieval of relevant resident information. - Ensure that existing user roles and authorization rules are respected by the AI. - Implement appropriate safeguards against unauthorized data disclosure and hallucinations. - Define appropriate AI logging, auditing and monitoring. - Support integration of the chatbot into the existing ION platform. - Document the architecture and security approach. What We Are Looking For Strong practical experience with: - LLM application development - Retrieval-Augmented Generation (RAG) - AI agents / tool calling - Vector databases and embeddings - Secure cloud architectures - Role-based access control (RBAC) - APIs and SaaS application integration - LLM evaluation and hallucination mitigation - Privacy and security for sensitive data Experience with healthcare, regulated environments, GDPR/Swiss data protection, or data-residency-sensitive AI architectures would be a significant advantage. Important We are not looking simply for someone who can connect our application to ChatGPT or another public LLM API. We need someone who understands the architecture, security and data-protection implications of using generative AI with sensitive resident information. When Applying, please briefly explain: - What architecture you would recommend for this use case. - How you would ensure that sensitive resident data remains within Switzerland. - Which LLM/model and hosting approach you would consider. - Your experience implementing RAG or AI assistants over private enterprise data. - Any previous experience with healthcare or other sensitive/regulated data. - Examples of similar AI solutions you have implemented. Please avoid generic AI-generated proposals. We are particularly interested in understanding how you would solve the Swiss data-residency requirement technically.