Staff Conversational Designer

This is a fully remote position, open to applicants in California.

📋 Description

• Define the persona and voice systems for Deepgram's voice agents across various experiences and use cases.

• Collaborate with ML and Engineering to design turn-taking, barge-in, and end-of-turn behaviors.

• Optimize responsiveness in relation to interruption risk for each specific use case.

• Create strategies for conversational repair, handling no-match and no-input scenarios, as well as confirmation techniques.

• Implement guardrails that necessitate confirmation prior to executing high-stakes actions.

• Manage latency-aware pacing and perceived responsiveness, including aspects such as brevity, backchanneling, hold speech, and filler speech.

• Develop evaluations for conversational quality and establish practices for transcript reviews.

• Transform production failures into a systematic design process.

• Build reference experiences for agents and provide design guidance aimed at developers for the Voice Agent API.

• Collaborate with ML and Research teams on ASR and TTS behaviors, along with criteria for conversation quality.

• Define principles for conversation design, review standards, and a shared vocabulary.

• Report directly to the Director of Product Design as a Senior individual contributor.

⛳️ Requirements

• Extensive experience in designing conversational behaviors for LLM-based voice agents or assistants.

• Proficient understanding of the speech pipeline from ASR through LLM to TTS.

• Solid knowledge of where design decisions are integrated within the speech pipeline.

• Proven track record in designing turn-taking, interruption, repair, and confirmation patterns that have been successfully implemented in production.

• Experience in developing quality measurement practices, including evaluations, transcript reviews, benchmarks, or structured failure-analysis loops.

• Exceptional writing skills, encompassing sample dialogues, design guidance, and comprehensive documentation.

• Ability to influence engineering and ML partners without direct authority.

• Experience in designing for developers or technical users, including APIs, SDKs, and documentation.

• Working knowledge of AI practices and insights on the appropriate use of AI tools.

• Availability during Pacific Time; preference for candidates residing in San Francisco.

• Must address U.S. location and U.S. work authorization/visa sponsorship inquiries.

🏝️ Benefits

• Equity

• Bonus

• Remote work options

• AI-first work environment

• Access to global investors and strategic partners (context for company compensation)