Director of Research, Text-to-Speech

COMPANY OVERVIEW Deepgram is bringing the world’s most natural interface, voice, to the technology we use every day. Our speech models and real-time infrastructure power AI systems that keep up with the way people actually speak. More than 5,000 organizations run conversational voice AI on Deepgram in production, including Twilio, Cloudflare, Sierra, Decagon, Vapi, Cresta, Granola, and Jack in the Box. The Deepgram Voice Agent Platform brings together speech-to-text, text-to-speech and speech-to-speech, with voice-native foundation models available through cloud APIs or self-hosted and on-premises deployments. Built for accuracy, low latency and cost efficiency, Deepgram makes high-quality voice experiences practical to build and run at scale. More than 650,000 developers have used Deepgram to process more than 200,000 years of audio. Every day, more of the world’s conversations with AI run on our models. Operating Principles: Deepgram is a research company that productizes as fast as possible. Four Operating Principles shape how we think, decide, and work together: 1. Compress before you communicate. You think it through first, then lead with the bottom line. 2. Resist entropy. You question what already exists before adding more process or tools. 3. Solve the category. You fix recurring problems at the source, not one case at a time. 4. Work at the frontier. You go after problems no one has solved yet. We look for people who already work this way, and these principles shape how we grow and evaluate our team. AI-FIRST BY DEFAULT We build AI that millions of people talk to, and we run the company the same way. AI is how we centralize work here, and is core to how we operate. That means more than using the tools. We expect you to: - Build, not just use. Design your own workflows and agents, and use AI to remove whole categories of work instead of doing the same work faster. - Own the output. AI can speed up the work. It doesn’t make anyone less accountable for it. You understand, verify, and stand behind everything you share, including what an agent produces for you. - Spend the time you get back at the frontier. The capacity AI creates goes toward problems no one has solved yet. - Keep pace. Models and tools change every month, and so will your work. You adopt what’s new quickly and drop what no longer earns its place. We judge AI by what it helps you deliver and whether you stand behind the result, not by how often you use it. In interviews, expect to walk us through a workflow you built and where you drew the line between AI and your own judgment. THE OPPORTUNITY We're looking for a Director of Research to own our Text-to-Speech program end to end — the research strategy, the technical bets, and the models that ship. This is a hands-on leadership role: you set direction and stay in the details, shaping architectures, experiments, training strategy, and evaluation. You'll also build the team and operating model that let exceptional researchers move fast without lowering the bar. The mandate is broad: advance the state of the art in speech generation, and turn those advances into models in production. WHAT YOU'LL DO - Own the TTS research and model roadmap — decide which technical directions can materially move speech-generation quality, including the expensive and non-obvious ones, and recognize when an approach should change or die. - Drive advances across neural audio modeling, prosody and expressiveness, controllability, multilingual speech, voice identity and consistency, data and training strategy, post-training, and inference performance — and make sure they land as measurable production gains. - Stay deeply technical: review research, challenge assumptions, design experiments, diagnose model failures, and take on the highest-leverage problems yourself. - Build evaluation and benchmarking that explains why models improve, not just whether they did — automated metrics alongside human perceptual assessment. - Lead a mix of individual contributors and tech lead managers. Hire and develop both, hold an exceptionally high technical bar, grow senior researchers into technical leaders, and set direction across sub-teams while pushing decisions down to the people closest to the work. - Partner with engineering and product leadership on ship-readiness, and represent Deepgram's TTS research internally and externally. YOU'LL LOVE THIS ROLE IF YOU - Want to own a frontier-model program end to end rather than optimize one slice of someone else's agenda. - Care about the breakthrough and the ship in equal measure. - Still love training runs, failed evals, and architecture arguments, even as most of your impact now comes through the people you lead. - Get energy from deciding which bets are worth making when the answer isn't obvious and the experiments are expensive. YOU MUST HAVE - Deep expertise in modern TTS, speech generation, or audio generative modeling, with a track record of personally training and improving large-scale neural models. - Command of the modern speech-generation stack and the open problems behind naturalness, expressiveness, controllability, robustness, voice consistency, and inference cost. - A history of setting research direction under genuine uncertainty: prioritizing experiments, allocating compute and researcher time, and killing approaches that aren't working. - Experience leading researchers and research engineers through other technical leaders — developing tech lead managers or equivalent, setting direction across sub-teams — while staying technically influential yourself. - AI as your default mode of work, not an occasional tool. You've rebuilt how you and your team operate around it, and you have a specific, earned view of what it still can't do in speech research. - The ability to make complex technical tradeoffs legible to product, engineering, and executive audiences. IT'D BE GREAT IF YOU HAD - TTS or generative-audio models deployed at meaningful production scale. - Built or substantially scaled a high-performing AI research organization. - Sophisticated evaluation systems for generative speech; expressive or multilingual generation, voice cloning and adaptation, or controllable generation. - Recognized external contributions — publications, open source, patents, invited talks — in speech synthesis, neural audio codecs, speech language models, or multimodal models. - Experience in fast-moving startup or research environments that routinely take models from idea to production. Notice: We're aware of individuals impersonating Deepgram recruiters. All legitimate Deepgram recruiting communication comes from an @deepgram.com http://deepgram.com email address. If you've received a message claiming to be Deepgram, please forward it to careers@deepgram.com.