AI Engineer (Voice)

We're looking for an AI Engineer to own HiKana's real-time voice conversation stack—speech recognition, TTS, LLM dialogue, and latency. HiKana is an iOS and Android app where English speakers practice Japanese by talking with Kana, an original AI anime character. You'll work closely with our technical co-founder in a small Tokyo team, shipping improvements that learners feel immediately in every conversation. Responsibilities Design and improve real-time voice AI: ASR, TTS, turn-taking, barge-in, and end-to-end latency Integrate and tune LLMs for spoken Japanese learning dialogue Own reliability and quality metrics for live voice sessions on iOS and Android Ship iteratively in a small, high-ownership team Requirements Strong English (required). Business-level Japanese is required (日本語ビジネスレベル必須). Genuine motivation to learn Japanese (required)—our product and users are built around that journey Hands-on experience with real-time voice or conversational AI Comfortable shipping production systems, not only prototypes Based in or willing to work from Tokyo (Shibuya) Nice to have Japanese ASR/TTS or multilingual voice models Mobile or realtime audio (WebRTC) experience Interest in language learning or anime/IP consumer apps Work style: This role is office-based at our Tokyo office (hybrid; full remote is not possible). Applicants must already live in Japan and hold a valid right to work in Japan, as we do not offer visa sponsorship for this position.