We develop human-centered speech and language technologies that can understand, generate, and protect the human voice. Our research focuses on expressive speech synthesis and voice conversion, large audio-language models, affective computing, speech deepfake detection and voice security, and speech and language technologies for health. Across these areas, we aim to build systems that capture not only what people say, but also how they say it, while remaining secure, trustworthy, and useful in real-world applications.