Senior Voice AI Engineer / Voice Platform Engineer

About the job Parse AI is hiring a Senior Voice AI Engineer / Voice Platform Engineer to help develop our scalable AI call center platform for hospitality operators. Important: To be considered for this role, please complete this short screening form: https://forms.gle/HJAZqcLczzcXaH8H6 Only candidates who submit the form will be reviewed. We are looking for a Senior Voice AI Engineer / Voice Platform Engineer to own the real-time voice layer of our platform. This includes telephony, streaming audio, STT, LLM orchestration, TTS, barge-in, latency, call transfers, multilingual voice, and production reliability. This is not a chatbot role. This is a production voice engineering role for someone who understands how real phone calls behave. Role Mission Own the real-time voice pipeline from phone and SIP ingress through speech recognition, turn detection, LLM orchestration, text-to-speech, interruption handling, and transfer flows. Your goal is to make AI phone calls feel natural, reliable, low-latency, and scalable across many restaurants, hotels, tenants, and call center environments. What You Will Work On Design and improve real-time STT → LLM → TTS voice pipelines Tune voice providers such as Deepgram, OpenAI, ElevenLabs, Cartesia, Hamsa, or similar tools Improve VAD, endpointing, silence handling, turn-taking, and barge-in behavior Build pronunciation and recognition lexicons for names, restaurant terms, hotel terms, accents, and multilingual use cases Work on Twilio Media Streams, SIP trunking, call transfers, warm transfers, and telephony reliability Measure and improve voice latency across STT, LLM, TTS, network, and telephony layers Debug production voice issues such as garbled audio, clipping, echo, slow responses, false interruptions, missed interruptions, and failed transfers Help scale concurrent call capacity across multiple customers and environments Partner with AI and backend engineers to make sure the agent experience is fast, safe, and production-ready What We Are Looking For We are looking for someone with real experience building and operating production voice systems. Strong candidates will have experience with: Real-time voice agents or voice platforms STT, TTS, and speech-to-speech systems Twilio Media Streams, SIP trunks, WebRTC, LiveKit, Telnyx, or similar systems Voice activity detection, endpointing, barge-in, and interruption handling Streaming audio, audio buffers, and telephony audio formats Voice latency measurement and optimization Production debugging of real phone call issues Multilingual or accent-heavy voice environments Provider evaluation and tuning across STT and TTS vendors Strong Plus Experience with restaurant, hotel, hospitality, or contact center voice agents Experience with LiveKit Agents, Rapida, Modal, or serverless voice runtimes Experience supporting 50+ concurrent calls Experience with Arabic, French, Polsih, MENA accents, or multilingual speech systems Experience building observability for voice quality, turn latency, interruptions, and transfers Experience building voice AI evals, simulations, and pre-production testing frameworks for latency, accuracy, barge-in, multilingual performance, and call readiness What This Role Is Not This is not a general LLM research role. This is not a frontend role. This is not generic backend CRUD work. This is not a simple voice API integration role. We need someone who can help make the actual phone call work in production. Ideal Candidate You are likely a strong fit if you: Think in milliseconds, audio formats, and real call behavior Have shipped real voice systems used by actual callers Can explain what broke in production and how you fixed it Understand that voice quality depends on telephony, STT, TTS, latency, interruption handling, and reliability working together Care deeply about making AI calls feel natural, fast, and dependable Location and Work Style This is a full-time remote role. Candidates should be able to overlap with MENA and Europe working hours, with occasional US timezone overlap when needed. We are especially interested in candidates who have shipped real production voice systems and can bring practical experience from live deployments.