Research focuses on machine learning for audio and speech, including generative modeling, controllable sound synthesis, multimodal language-audio representation, neural audio codecs, fine-grained editing, and the robustness, ethics, and integration of AI tools for creative workflows.