Speech Processing ML Algorithm Engineer

What you'll do

• Design, train, and evaluate machine learning models for speech enhancement tasks like noise suppression and de-reverberation.
• Develop multi-microphone processing algorithms that meet low-latency and real-time hardware constraints.
• Implement streaming and causal processing techniques to balance quality, complexity, and latency trade-offs.
• Apply signal processing fundamentals, such as STFT analysis and synthesis, to audio data.
• Conduct speech quality evaluations using objective metrics like PESQ and STOI alongside subjective listening tests.
• Develop model efficiency techniques including quantization-aware training or distillation for production deployment.
• Research and reproduce state-of-the-art literature to improve existing speech processing technologies.
• Manage the full development lifecycle from early prototypes through to final product integration.

What we're looking for

• MS or PhD in a computational science field or 3+ years of experience in ML for speech processing.
• Deep understanding of transformers, recurrent neural networks, and convolutional neural networks for audio research.
• Experience designing, training, and evaluating models for noise suppression, dereverberation, source separation, or echo suppression.
• Experience building models that meet low-latency, real-time requirements including streaming and causal processing.
• Strong grounding in audio and speech signal processing fundamentals such as STFT analysis/synthesis.
• Proficiency with PyTorch and Bash, including version control, code review, testing, and reproducible experiments.
• Working knowledge of objective metrics (PESQ, STOI, SI-SDR, DNSMOS) and subjective listening tests for speech quality evaluation.
• Familiarity with model efficiency techniques like quantization-aware training or distillation and high-performance cloud computing.