Description
Summary LLM Engineer — Simulated Primary Care Patient Interface. Please note this contract is only open to freelancers located in the EU. We are looking for an experienced LLM engineer / conversational AI specialist to help us complete the LLM layer of a high-fidelity sandbox version of our healthcare platform. This is a simulation/demo environment used with pilot users and design partners. The role is focused on making the patient-facing LLM interface work convincingly in a simulated primary care setting, where users interact with the system to complete flows such as appointment booking, patient access requests, consent capture, and guided intake. The product is not being positioned as a medical device. The focus is on building a realistic, safe, structured simulation of how an AI-supported primary care access interface should behave. What we need built: We require a AI dev to lead the build of the LLM-powered patient interface so it can: Handle realistic patient conversations in a simulated primary care environment Guide users through appointment booking and access workflows Collect and structure relevant patient information Mimic appropriate guardrails for healthcare-facing conversations Support integrated consent flows Route different types of user intent appropriately Avoid unsafe, overconfident, or clinically inappropriate responses Produce structured outputs that can be used elsewhere in the sandbox The right person will have experience with Building production-quality LLM interfaces, not just basic chatbots Prompt architecture and conversation design Multi-turn dialogue flows Guardrails, refusal handling, escalation logic, and safety boundaries Structured outputs from LLM conversations Intent detection and routing Simulated user journeys or sandbox/demo environments Testing LLM behaviour across edge cases Designing AI experiences where accuracy, tone, and user trust matter Responsibilities Design and implement the LLM behaviour for the patient-facing interface Define the conversation logic, prompts, routing, and safety rules Build flows for simulated patient access, appointment booking, consent, and intake Ensure the LLM behaves consistently across common and edge-case scenarios Create test cases and evaluation examples for expected behaviour Work with our existing product team to integrate the LLM layer into the sandbox How to apply: A short summary of relevant LLM products or interfaces you have built Examples of work involving multi-turn conversational flows Any experience with safety, guardrails, structured outputs, or intent routing We are especially interested in people who have built LLM interfaces for healthcare environments where the AI has to behave predictably and safely.