ML Engineer - Voice AI

Role overview Develop and deploy machine learning for real-time voice applications, combining speech processing with production engineering. The work includes detecting answering machines, analyzing speech and sentiment, evaluating call quality, and integrating models into a live dialing pipeline.

Responsibilities - Train and deploy answering machine detection models, targeting p99 latency below 3,500 milliseconds. - Build real-time transcription and sentiment analysis pipelines. - Develop call-quality scoring models using measures such as MOS and jitter. - Optimize model inference for edge deployment on serverless functions. - Work with the platform team to integrate ML capabilities into the dialing pipeline.

Requirements - At least three years of ML engineering experience, including production model deployment. - Experience processing speech or audio, such as ASR, TTS, or speaker diarization. - Proficiency in Python and familiarity with TypeScript. - Knowledge of model serving tools or formats such as ONNX or TensorRT.

Nice to have - Published research or substantial open-source contributions.

Benefits and work setup - Full-time, remote role open to candidates globally.