Founding AI Product Engineer

I am looking for a strong AI/full-stack engineer with excellent product sense to help build an MVP for a new personal-intelligence product. For many years I have studied several systems that use birth information to identify patterns in personality, behavior, relationships, motivation, decision-making, strengths and blind spots. I have accumulated substantial source material, case studies, observations, hypotheses and data. I now want to turn this knowledge into a modern AI product. Users would create profiles for themselves or other people and use the product to explore questions such as: What is distinctive about this person? What motivates them? What are their strengths and blind spots? How do they make decisions? How do they communicate? What environments bring out their best or worst? How are two people similar or different? What should this person understand about themselves? The product should feel sophisticated and useful — not like a traditional astrology or numerology website. What I need I am not looking for an order-taking developer. I am looking for someone who can work closely with a founder/domain expert, absorb a large amount of unstructured knowledge, help structure it, make strong technical and product decisions, and build a compelling MVP. You should be strong in: LLM applications RAG / retrieval Structured knowledge architecture Full-stack development React / TypeScript APIs and backend architecture AI evaluations Product analytics Rapid prototyping MVP product design Strong product sense is extremely important. I want someone who thinks about: What belongs in Version 1 What should NOT be built yet How to create a strong first-user experience What makes the product meaningfully different from ChatGPT What creates a “wow” moment What drives repeat usage How to measure whether users find outputs accurate and useful How to use analytics and feedback to rapidly improve the MVP Current architecture thinking My current lean-stack idea is approximately: React / TypeScript → Cloudflare → Railway → Supabase / Postgres / Auth / Storage / pgvector → OpenAI and/or Anthropic → PostHog → Stripe → GitHub I am NOT committed to this architecture. I want someone capable of challenging it and recommending a simpler or better solution where appropriate. The underlying system will likely separate: Deterministic birth-related calculations Structured proprietary knowledge Retrieval / RAG LLM interpretation and conversation I do not want an architecture where the LLM simply receives a birthday and invents an interpretation. Claude Code I strongly prefer someone who actively uses Claude Code or similar AI coding tools and understands how to use them to dramatically accelerate development while still maintaining strong engineering judgment. Initial project The first engagement would be approximately 4–6 weeks. The goal is to: Understand and structure the core knowledge Design the MVP architecture Build the calculation / profile layer Build an initial knowledge and retrieval system Create a compelling self-profile experience Allow profiles of other people Add conversational exploration Add authentication and saved profiles Add analytics and feedback Launch a private beta with approximately 20–50 users Measure usage and iterate The goal is NOT to build everything. The goal is to find the smallest version users find unusually compelling. About me I will provide the domain expertise, source material, frameworks, cases, hypotheses, data and judgment about output quality. I specifically do not want to spend my time coding, cleaning data, manually organizing knowledge or managing multiple developers. Ideally, we have one or two structured working sessions each week and you own execution between them. When applying Please answer these questions briefly: What stack would you use for this MVP, and what would you change about the stack I described above? How would you separate deterministic calculations, structured knowledge/RAG and LLM reasoning? Do you actively use Claude Code? How? If we had six weeks, what would you put in the MVP and what would you deliberately leave out? What would you do to create a genuine “wow” moment for a first-time user? How would you measure whether users actually find the AI's insights accurate and useful? Please link to one AI/LLM product you personally built and shipped, and briefly describe your role. Please do not send a generic proposal. I am much more interested in how you think about the product than in a long résumé. I expect to begin with a small paid trial before committing to a longer engagement.