We're looking for an experienced prompt engineer to help us improve the output quality of an AI accountability coach feature inside our digital wellness iOS app. Currently our output is not hitting the tone, depth, or personalization level we need. The Project Our AI coach helps users reflect on their phone habits, set intentions, and stay accountable. We need help with one specific prompt that powers a key output on our home screen. Right now, the responses feel too generic and don't consistently match the voice and therapeutic style we're going for. We need someone who can evaluate what we have, identify gaps, and re-engineer this single prompt to produce responses that feel warm, insightful, and genuinely helpful. If this initial engagement goes well, we have additional prompts across the app that we'd love to collaborate on together. What We're Looking For Proven experience with prompt engineering for consumer-facing AI products Familiarity with LLM API (strongly preferred) Understanding of system prompts, few-shot examples, chain-of-thought techniques, and structured output control Bonus: background in psychology, coaching, or behavioral health content Ability to iterate quickly based on feedback and test outputs Deliverables Audit of the current prompt with specific recommendations A revised, production-ready prompt A testing framework or evaluation criteria so we can measure improvement Documentation of the prompt logic so our team can maintain and iterate going forward To Apply Share examples of prompt engineering work you've done, especially for conversational AI or coaching/wellness use cases. Describe your process for diagnosing and fixing underperforming prompts. If you have experience with the Claude API specifically, please highlight that.