Research and develop AI-powered audio source separation and extraction algorithms for speech and real-world sounds. Train and evaluate PyTorch models using microphone arrays, spatial cues, and auxiliary sensors; address noise, reverberation, overlapping speakers, latency, memory, and computational constraints. Collaborate with DSP and embedded software teams to deploy real-time models on physical devices, and contribute to patents and research publications.

Research and develop novel AI-powered audio understanding algorithms, including sound event detection, localization, acoustic scene analysis, anomaly detection, and multimodal context understanding. Build datasets, pipelines, evaluation protocols, and real-time prototypes for embedded devices. Collaborate with machine learning, DSP, psychoacoustics, and software engineering teams to integrate models into product platforms and contribute to publications or patents.

Develop agentic workflows that automate audio ML experimentation, including data preparation, training, evaluation, benchmarking, regression testing, reporting, and reproducible execution. Build tooling for GPU scheduling, experiment tracking, artifact management, and resource monitoring across on-premises and cloud infrastructure. Partner with audio researchers to create reliable, observable systems and dashboards for real-time audio AI development.