What you’ll do
- Develop realistic, human-sounding AI voices with natural emotions, accents, intonation, and expressions
- Build and fine-tune voice cloning, Text-to-Speech (TTS), and Speech-to-Text (STT) models
- Develop multilingual voice capabilities, including English, Arabic, Hindi, and regional languages
- Optimize real-time voice conversations for low latency, natural turn-taking, and interruptions
- Research, customize, and deploy open-source speech AI models for cloud and offline environments
- Integrate speech models with LLMs and conversational AI platforms
What we’re looking for
- 2–7 years of experience
- Strong Python, PyTorch, and Hugging Face experience
- Solid understanding of STT, TTS, NLP, NLU, speech processing, and deep learning
- Hands-on experience with models such as Whisper, XTTS, F5-TTS, VITS, Fish Speech, or similar
- Knowledge of audio processing, phonetics, prosody, speaker embeddings, and neural vocoders
- Experience with model fine-tuning, GPU optimization, and real-time audio streaming
Nice to have
- Familiarity with WebRTC, WebSockets, FastAPI, and Docker
- Demonstrable voice cloning, natural speech synthesis, or speech AI projects (share your GitHub profile and voice AI demos when you apply)