Research & Technical Intern, NiTHUB — University of Lagos (UniLag) (Yoruba POS Tagger)
Worked on development of a Yoruba NLP pipeline for a low-resource part-of-speech tagging task, which required building and validating annotated training data for the model. Contributed to data annotation and dataset preparation steps and supported evaluation methodology for morphosyntactic tagging under limited computational resources. The role focused on obtaining labeled examples appropriate for POS prediction in an African language context. • Supported data annotation workflows and quality checks for Yoruba linguistic items • Informed model architecture and evaluation decisions based on labeled data behavior • Assisted end-to-end cycles including data collection/annotation, training, and evaluation • Collaborated with research teams to improve language-technology outcomes for low-resource NLP