Director of Research, Text-to-Speech at Deepgram

Role overview

This job description has been rewritten in fresh language using the facts in the listing. Source credit and original listing details are available below.

Deepgram is hiring a Director of Research to lead its Text-to-Speech (TTS) program from research strategy through production models. The role combines hands-on technical leadership—shaping architectures, experiments, training plans, and evaluation—with building and guiding the team and operating practices needed to move quickly while maintaining a high technical standard.

The director will set the TTS roadmap, choose research priorities, and translate advances in speech generation into measurable production improvements. The work spans neural audio modeling, prosody and expressiveness, controllability, multilingual speech, voice identity and consistency, data and training, post-training, and inference performance. The position also partners with engineering and product leaders on readiness to ship and represents Deepgram’s TTS research internally and externally.

Deepgram describes its workplace as AI-first: employees are expected to regularly use and experiment with advanced AI tools, integrate them into their work, and adapt as the technology changes. The role is remote in the United States, full-time, with annual compensation of $288,000–$361,000.

Role at a glance: Deepgram is seeking a remote U.S. Director of Research to lead its Text-to-Speech research, team, and production roadmap.

Responsibilities mentioned in the listing

• Set and own the TTS research and model roadmap, selecting technical directions and revising or ending approaches that are not working.
• Advance speech-generation quality across neural audio modeling, prosody, expressiveness, controllability, multilingual speech, voice consistency, data and training, post-training, and inference performance.
• Review research, test assumptions, design experiments, diagnose model failures, and tackle high-impact technical problems.
• Develop evaluation and benchmarking that combines automated measures with human perceptual assessment and explains model changes.
• Lead individual contributors and tech lead managers; hire and develop researchers, cultivate technical leaders, set direction across sub-teams, and delegate decisions to those closest to the work.
• Work with engineering and product leadership on whether models are ready to ship, and represent TTS research internally and externally.

Qualifications stated in the listing

• Deep expertise in modern TTS, speech generation, or audio generative modeling, including a record of personally training and improving large-scale neural models.
• Strong command of current speech-generation methods and challenges in naturalness, expressiveness, controllability, robustness, voice consistency, and inference cost.
• Experience setting research direction amid uncertainty, prioritizing experiments, allocating compute and researcher time, and stopping ineffective approaches.
• Experience leading researchers and research engineers through technical leaders, developing tech lead managers or equivalent, and guiding multiple sub-teams while remaining technically influential.
• An AI-first approach to work, including adapting personal and team workflows around AI tools and a well-founded perspective on their limitations in speech research.
• Ability to explain complex technical tradeoffs to product, engineering, and executive audiences.
• Preferred: TTS or generative-audio models deployed at meaningful production scale.
• Preferred: Building or substantially scaling a high-performing AI research organization.