Deep Learning Research Engineer - Speech & Audio

About the opportunity

Develop state-of-the-art neural speech synthesis, multi-speaker diarization, and streaming voice translation models.

What you will do

• check_circle Train modern autoregressive and diffusion models for high-fidelity zero-shot expressive voice synthesis.
• check_circle Build streaming end-to-end automatic speech recognition (ASR) pipelines with sub-200ms latency.
• check_circle Implement noise-robust audio feature extractors and self-supervised speech representations (wav2vec, Whisper variants).
• check_circle Deploy optimized ONNX and TensorRT speech inference engines to Azure cloud endpoints.

What we are looking for

• arrow_circle_right MS or PhD in Computer Science, Electrical Engineering, or equivalent practical experience.
• arrow_circle_right 4+ years training deep learning models specifically for audio processing, speech recognition, or acoustics.
• arrow_circle_right Strong proficiency with PyTorch, ONNX, C++, and Python audio DSP libraries (torchaudio, librosa).

Skills & Tech Stack