Ayanda Goboza

I've supported Appen, TransPerfect, Smartcat, and Bilingoplus as a freelance AI linguist and data specialist since 2016, curating multilingual datasets for LLM fine-tuning and model alignment across legal, fintech, energy, and streaming work.

I evaluate model responses for hallucinations, semantic drift, and bias, maintaining 98%+ quality ratings. My linguistic error reports helped improve Dutch-to-English model fluency by 12%, while my Python preprocessing removed syntax, formatting, and duplicate-segment issues before annotation.

I also resolve localization edge cases, including 200+ contextual token mismatches, and collaborate asynchronously with engineers and linguists on EMEA streaming metadata deployments. Earlier, I built data-capture workflows that reduced reporting errors by 15% and processed public-sector datasets with 0% data loss.