I evaluate AI-generated responses for accuracy, relevance, and helpfulness, comparing them with instructions and established quality guidelines.
For AI training, I annotate and categorize text, identify factual errors and inconsistencies, and provide human feedback to improve language model performance.
I also transcribe English and Swahili audio, review transcripts for accuracy, and identify speakers and segment recordings.

