Staff Product Designer, Conversational AI

You will define and own the conversational behavior of Deepgram's voice agents, including persona, turn-taking, and repair strategies. You will also partner with engineering and ML teams to build evaluation frameworks and developer-facing guidance for voice agent APIs.

Responsibilities

• Define persona and voice system for Deepgram voice agents, ensuring coherence across use cases
• Design turn‑taking, barge‑in, and end‑of‑turn behavior with ML and engineering, balancing responsiveness and user interruption risk
• Create conversational repair, no‑match, no‑input handling, and confirmation strategies with guardrails for high‑stakes actions
• Own latency‑aware pacing, back‑channeling, filler speech, and perceived responsiveness within real‑time budgets
• Establish conversational‑quality evaluation frameworks and transcript‑review practices to turn failures into repeatable design loops
• Build reference agent experiences and developer‑facing design guidance showcasing best‑practice conversations on the Voice Agent API
• Partner with ML and research on ASR and TTS behavior and define quality criteria for good conversations
• Set conversation‑design principles, review standards, and shared vocabulary for the broader team

Requirements

• Deep experience designing conversational behavior for LLM‑based voice agents or assistants
• Fluency with the speech pipeline (ASR → LLM → TTS) and where design decisions reside
• Proven track record of designing turn‑taking, interruption, repair, and confirmation patterns that ship in production
• Experience building quality measurement systems such as evals, transcript reviews, or benchmark loops
• Exceptional writing craft for sample dialogs, design guidance, and documentation
• Ability to influence engineering and ML partners without direct authority
• Experience designing for developers or technical users (APIs, SDKs, documentation)
• Working AI practice with insight into appropriate tool usage
• Bonus: experience with wizard‑of‑Oz testing, eval tooling for LLM or voice quality, high‑stakes conversation domains, multilingual design, or B2B growth environments

Key Skills