Speech & audio ML, voice embeddings/representation learning, personalization + recommender systems, time-series modeling, human-in-the-loop evaluation, privacy-preserving ML, responsible AI, and MLOps for real-world deployment.
No public activity