Chinedu Charles Iloanya

I've evaluated AI-generated responses at Outlier AI's Aether Project, assessing accuracy, clarity, and user experience while providing structured feedback to improve outputs.

Across OneForma's Lighthouse Project, I've annotated and reviewed datasets, transcribed audio, and reviewed Japanese-English outputs for accuracy, clarity, and consistency. I work carefully within detailed guidelines, quality benchmarks, and deadlines.

At LinguaData Solutions, I've segmented and annotated more than 400 hours of natural conversational English in Praat, producing time-aligned transcriptions and prosodic boundaries. I validate automated speech boundaries through spectrogram and waveform analysis, use ToBI conventions, and preserve disfluencies, overlaps, false starts, and non-linguistic sounds in verbatim transcription.

My background combines linguistic analysis, data quality work, and independent remote delivery. I bring strong editorial judgment, critical thinking, and rigorous attention to detail to AI training and language-data projects.