Job Description - Member of Technical Staff (Speech)
Location: San Francisco (on-site at our offices)
The Opportunity
Speech is becoming AI's next interface: text to speech, speech to text, voice cloning and real-time voice agents are moving from demos into infrastructure, and our speech benchmarks are how the industry tracks who is winning. We're hiring into our speech pillar to drive that coverage.
You'll build and extend our speech evaluations and arenas, from streaming speech to text and voice cloning to next-generation speech-to-speech intelligence with benchmarks like AgentTalk, and work at a deep technical level with the leading speech AI companies to benchmark their latest models as they launch.
What You'll Do
• Benchmark the Speech Frontier: Own coverage across text to speech, speech to text and speech to speech, benchmarking new models and providers as they launch
• Design Speech Evaluations: Build and extend the evaluation frameworks, prompt libraries and arenas that reflect how developers and creators actually use speech models, including agentic voice through AgentTalk
• Partner with Speech Leaders: Work with the top speech AI companies in the world at a deep technical level to benchmark their models and shape how the industry measures voice
• Publish Influential Analysis: Produce the leaderboards, reports and analysis that shape how the industry understands speech AI progress
• Drive Product Direction: Shape the roadmap of our speech benchmarking platform together with our engineers and pillar lead
• Become AI-Native: Embrace an AI-native workflow, using cutting-edge AI tools to generate leverage in a fast-changing industry and maintain our competitive edge in AI benchmarking
What We're Looking For
You should know speech AI from the inside.
Backgrounds include: product, research or engineering roles at speech AI companies (e.g. ElevenLabs, Cartesia, Inworld, Sesame, Deepgram, AssemblyAI, Rime or similar), voice teams at larger platforms (e.g. OpenAI, Google, Microsoft), or teams building products on text to speech, speech to text or real-time voice.
Required:
• 3+ years of professional experience, including at least 1 year working hands-on with speech AI
• Strong analytical and critical thinking skills
• Proficiency in Python and data analysis
• Hands-on familiarity with modern speech models and their evaluation: quality assessment, word error rate and latency measurement, and preference testing
• Genuine, demonstrable interest and knowledge of Frontier AI. We want people who have informed opinions about where AI is heading, not just people who use AI tools
• Competitive compensation including equity
1
