Founding AI Voice Engineer

About this opportunity

[TW] Every 40 seconds, someone in the world dies by suicide. But long before that moment, there are often hundreds of smaller moments. A night when you can't sleep. A breakup. Anxiety before work. The feeling that you're losing control. Or simply a moment when you need someone to talk to and nobody is there. A therapist might be available in 6 month. A friend might be asleep. And sometimes, you don't want to talk to either. We believe there should always be someone there. Someone you can talk to at 2:37 in the morning. That's what we're building. A voice-first AI companion for mental health and personal development, available to anyone, anywhere. And we're looking for a Founding AI Voice Engineer to build it with us from zero. What you own You own the technology behind the product end-to-end: Build and ship iOS & Android apps Build real-time voice-to-voice AI agents for affective full duplex dialogue Select and integrate the right LLMs and voice models Build RAG, memory and context systems Develop AI personas, behavior and conversational logic Implement guidelines, guardrails and safety systems Translate psychological frameworks into AI behavior together with experts Build evaluation, monitoring and testing systems Own backend, infrastructure and deployment Optimize latency, quality, reliability and inference cost Take the product from prototype → production → scale Experience with clinical AI, mental health or digital health is not required, but a strong advantage. Who we're looking for You've done this before. We're not looking for someone who wants to get into Voice AI. We're looking for someone who has already: Built real-time speech-to-speech / voice-to-voice AI agents Deployed them into real-world production environments Worked hands-on with STT, LLM and TTS pipelines and/or native speech-to-speech models Solved latency, interruptions, turn-taking and streaming Built agents capable of handling long, dynamic conversations Worked with LLM orchestration, RAG, memory and context management Owned systems beyond the prototype — including deployment, monitoring, reliability and scale You understand the difference between a voice demo that works for five minutes and a production voice agent people can actually use every day. How we work This is early stage. No engineering team behind you. No product manager writing tickets. No six-week planning cycles. Plan in the morning. ship until the evening. We want someone who: Works extremely hard Moves extremely fast Operates independently Takes full ownership Figures things out without waiting for instructions Can go from idea to production Cares deeply about what we're building You will have enormous freedom and enormous responsibility. The mission We are not trying to replace therapists. We want to make mental support and personal development available to anyone with a smartphone, regardless of geography, income or time of day. If we get this right, we can build something millions of people turn to when they need mental support and advice. If that's something you want to spend the next few years building, we want to hear from you.