Senior AI Engineer (Voice Systems)
Core
Technical anchor for next-generation real-time voice bot systems, owning the intelligence layer from acoustic understanding to natural speech synthesis.
Role type
Senior IC machine-learning engineer (voice systems)
Builds
End-to-end voice pipelines (ASR, NLU, dialogue management, TTS) for real-time customer service interactions
Domain
Telecommunications / Voice AI / Natural Language Processing
Deliverable
production ML models
Required skills
ASR fine-tuning, TTS model training, speaker diarization, VAD, deep learning frameworks (PyTorch/TensorFlow), Python, model evaluation metrics (WER, RTF, MOS), LLM prompt engineering for voice, inference optimization (quantization, distillation)
Preferred skills
Multilingual/low-resource language ASR/TTS, telephony audio constraints (G.711 codecs), streaming ASR, open-source speech project contributions
Technologies
Whisper, Gipformer, wav2vec 2.0, F5-TTS, VoxCPM2, VITS, ESPnet, SpeechBrain, NeMo, Hugging Face, MLflow, W&B, Docker, ONNX, TorchServe, Triton
Responsibilities
Design and develop end-to-end voice pipelines; Fine-tune ASR and TTS models on domain-specific Vietnamese data; Build evaluation frameworks for voice AI quality; Optimize inference latency for real-time constraints; Develop intent classification and entity extraction components; Collaborate on model serving architecture and monitoring pipelines
Seniority
Senior, hands-on IC
