Yahia Kedadria

I’m a trilingual English language specialist with an academic foundation in linguistics, translation methodology, and cross-cultural communication. I bring native Arabic fluency and deep analytical command of English to create training and evaluation that reflects real-world language complexity.

Since 2022, I’ve worked as an AI Language Evaluator & Data Specialist on remote contract platforms including Appen, Remotasks, and OneForma. I stress-test advanced LLMs on English grammar, verb aspect, idiomatic usage, and cultural pragmatics, then document reproducible failure modes such as tense confusion, semantic drift, and pragmatic inappropriateness.

I evaluate model-generated text against factual accuracy and logical soundness benchmarks, flagging hallucinations and reasoning errors with structured error traces. I also apply contrastive analysis across Arabic–English and French–English to surface cross-linguistic transfer errors that monolingual evaluators often miss.

Alongside AI evaluation, I’ve worked as a freelance translator & linguist since 2021, delivering Arabic ↔ English and French ↔ English translations across technical, legal, and general-purpose domains. I onboarded with Midlocalize as a vendor linguist and use systematic quality-check procedures to verify accuracy, fluency, and cultural appropriateness—an approach I bring to my AI training work as well.