Abeera Najam

I build open-source infrastructure for measuring whether AI safety survives linguistic distribution shift. My work sits at the intersection of AI safety, multilingual NLP, evaluation, and AI governance, beginning with Urdu and other underrepresented linguistic settings.

VERIAUDIT started from a simple observation: almost all public AI safety evaluation happens in English, while a large share of the world's AI users interact with these systems in low-resource languages, transliterated scripts, and code-switched speech. Measuring that gap — rather than assuming it doesn't exist — is the project's starting point.