Independent AI Data Contributor (Various Platforms) | AI transcription and labeling for Swahili audio
Provided transcription and linguistic labeling for complex Swahili audio to ensure accurate phonetic and intent capture. Labeled non-speech audio events such as background noise, overlapping dialogue, and audio reflections. Completed quality assurance against project guidelines to maintain high-fidelity datasets for ML. • Transcribe Swahili speech with attention to phonetics and speaker intent • Tag non-speech events (background noise, overlap, reflections) • Meet/exceed QA benchmarks for dataset accuracy • Troubleshoot crowdsourced annotation platform workflows