Welcome! 😎
Research
I work on multilingual speech and language technology, with a focus on resource-constrained languages and real-world speech. My research explores how we can build, evaluate, and improve models when data, computation, and linguistic resources are limited.
My recent work focuses on speech model evaluation, error diagnosis, data-centric methods, and,efficient multilingual models, particularly for African languages and accents. I am interested in understanding not only how well a model performs, but where it fails, why it fails, and how those failure patterns can guide more targeted use of available data.
More broadly, I am interested in building speech and language systems that are robust to the diversity and imperfections of real-world data without relying solely on larger models or larger datasets.