About the Role
Eunoie's Nora persona is built on a layered conversation architecture — not a single LLM call. You will design and own the full dialogue management stack: intent routing, state tracking, tone modulation, and safe-exit flows for distressed users.
What you'll do
- ›Build and maintain stateful multi-turn conversation engines with emotional velocity tracking
- ›Implement tone classifier pipelines that route between warmth, wit, and structured-support modes
- ›Design the Acknowledgment Gate layer that prevents hollow AI responses
- ›Integrate conversation memory with the emotional graph (velocity + acceleration modeling)
- ›Write and maintain gold-standard conversation evaluation suites
- ›Work with the product team to define conversation UX contracts and escalation protocols
What we're looking for
- ›Strong background in dialogue systems, conversational AI, or NLP-driven chat products
- ›Experience with LLM orchestration frameworks (LangGraph, LangChain, or equivalent)
- ›Ability to design evaluation harnesses for conversation quality — not just accuracy metrics
- ›Familiarity with safe messaging guidelines and distress-detection in text
- ›Experience deploying to real users with measurable conversation quality outcomes
- ›A GitHub showing real projects, not toy demos