At RWS TrainAI, I contribute to LLM training and evaluation projects, including response ranking, preference testing, safety assessment, and instruction-following analysis. I write structured justifications for scoring decisions within detailed project guidelines.
Across the Metis Project and OneForma, I evaluate multilingual responses and Turkish–English translations for relevance, reasoning, fluency, and accuracy. My sociology background informs my attention to analytical reasoning and bias, while my work also includes AI data annotation, running PubKA, and freelance 3D and graphic design.
