Conversational AI Engineer, System Prompt

Known • San Francisco, CA

2d • $170,000 - $220,000 • Onsite

About The Position

We’re looking for founding Conversational AI Engineers to build the prompt systems powering our voice-led onboarding and user experiences.

This is a unique opportunity to work with a hyper-personalized data-set, combining voice transcripts, images, and structured user data to empower real-time, personalized AI voice-led conversations at scale. You’ll work directly with Chen Peng, former head of ML at Uber Eats and Faire.

Requirements

• 2-3 Years in Conversational AI/NLP: Proven experience designing, testing, and deploying complex LLM applications and system prompts in high-traffic production environments.
• The Prompt Stack: Deep familiarity with state-of-the-art prompt engineering techniques (e.g., Few-Shot, Chain-of-Thought, ReAct).
• Agentic & RAG Architectures: Experience building the "brain" logic for LLMs using frameworks like LangGraph, LlamaIndex, or Haystack to manage complex, non-linear dialogue and dynamic knowledge retrieval.
• Production Hardened: You treat prompts as an engineering problem. You’ve optimized prompt systems for scale, API cost, and speed. You're comfortable with prompt version control, programmatic prompt optimization (e.g., DSPy), and building continuous integration pipelines for AI evals.

Responsibilities

• Prompt Orchestration & Context Optimization: Architecting the core system prompts and managing context windows to ensure highly responsive, contextually relevant, and logically sound AI reasoning without bloating token counts or causing latency spikes.
• EQ & Semantic Memory: Building prompt systems that allow Known to maintain a consistent, empathetic, and uniquely "Known" personality. You'll design mechanisms to seamlessly weave long-term user memories and preferences into real-time dialogue, while helping the user drive the conversation.
• Conversational Intelligence: Designing advanced prompt chains (and fallback logic) to gracefully handle conversational tangents, user interruptions, semantic end-of-turn conversation logic,, and complex emotional states so Known feels empathetic and responsive.
• Agentic Workflow Design: Implementing and maintaining the prompt-driven logic for multi-agent frameworks, where your system instructions act as the routing engine between the user, external APIs, and our internal matchmaking engine.
• Evals for Conversational Quality: Developing custom evaluation frameworks to measure "conversational success." You'll go beyond basic fact-checking to rigorously assess conversational dynamism, warmth, engagement, and hallucination reduction.

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level
• Education Level: No Education Listed