carlshizi/yoruba-speech-pipelines

A research writing sample proposing a data-centric speech pipeline for low-resource Yoruba ASR, TTS, and tone-aware evaluation. ## Building Low-Resource Speech Pipelines for Yoruba Using Data-Centric Methods This repository contains a methodological research paper that outlines a data-centric speech processing pipeline for Yoruba, a low-resource tonal language. The paper describes system design, data collection methodology, and an evaluation framework for Automatic Speech Recognition (ASR), Text-to-Speech (TTS), and tone-aware assessment. Empirical results are planned as future work, and this document is intended as a research writing sample reflecting my research direction.