Build and fine-tune speech and audio ML models (ASR, diarization, audio classification), create data pipelines for large audio datasets, design experiments and benchmarks, and deploy scalable training and inference workflows integrating voice capabilities into real-time applications.