About the Role
You will own the design and deployment of generative AI systems at the core of Eunoie's emotional intelligence platform. This is not a wrapper-on-top-of-GPT role — you will work directly on prompt architecture, retrieval-augmented generation, and multi-turn emotional reasoning systems.
What you'll do
- ›Design and implement production-grade generative pipelines using LLMs (Claude, GPT-4, Llama)
- ›Build and maintain a multi-pipeline RAG vector store with emotional context tagging
- ›Architect the DCRI Gate system (Detect → Clarify → Reframe → Insight) at the prompt layer
- ›Evaluate model outputs against emotional gold-standard conversation benchmarks
- ›Collaborate with the ML Research team to integrate fine-tuned emotion classifiers into the generation loop
- ›Instrument pipelines with latency, quality, and safety observability
What we're looking for
- ›Shipped at least one LLM-powered feature to real users in production — not a demo
- ›Deep familiarity with RAG patterns: chunking, embeddings, hybrid search, reranking, vector databases
- ›Experience with prompt engineering at scale — system prompts, few-shot chains, tool-use
- ›Comfortable with Python async, FastAPI or similar for serving inference endpoints
- ›Understanding of emotional AI, affective computing, or conversational UX is a strong plus
- ›Open-source contributions or a solid GitHub portfolio