Abdessamade Hadoudouni

I’m a bilingual AI evaluator bridging the gap between complex linguistic AI training and technical logic. I adhere to highly restrictive formatting guidelines, QA protocols, and verbatim transcription standards while assessing LLM responses in Arabic and English for factual accuracy, tonal consistency, and logical flow.

Since 2023, I independently train and evaluate AI by designing complex, multi-turn prompts to stress-test models on edge cases, cultural nuances (Maghrebi Arabic), and specific constraints. I also perform technical QA of AI-generated code (PHP, WordPress, CSS), identifying backend routing errors, database logic flaws, and inefficient execution, and I apply RLHF-style reinforcement learning from human feedback, including response ranking and hallucination detection.