Gabi Pelli

Evaluated high-difficulty prompts and AI-generated outputs for reasoning quality, conceptual consistency, instruction-following, nuance, and response robustness across multi-turn tasks. Identified failure patterns (ambiguity, shallow reasoning, conceptual collapse, framing errors) and performed PT-BR/English linguistic QA.

Evaluated AI-generated outputs for clarity, coherence, factual consistency, tone, reasoning quality, instruction-following, and audience fit. Compared model responses across writing, summarization, translation, reasoning, safety, and multimodal tasks, and refined prompts and guidelines for PT-BR/English quality criteria.

Designed and refined prompts, AI-assisted workflows, and evaluation frameworks to improve communication, content creation, and output quality. Reviewed AI-generated content for clarity, coherence, tone, structure, naturalness, reasoning quality, and audience alignment, including building PT-BR language quality frameworks.

Led content development and language adaptation for the HEADSCON ES launch campaign, using brandbook and persona definitions to craft more locally resonant messaging. Produced and maintained consistency across multi-format deliverables including social posts, video scripts, radio copy, and institutional texts.

Led senior copywriting and strategic communication projects for a large Brazilian financial cooperative system. Developed a nationwide verbal identity manual and created multi-format institutional and educational content, supporting executive and international-facing communication through interpretation and high-context language adaptation.

Graduate-level study in cognitive neuroscience and behavior, with interests in language, perception, reasoning, meaning-making, and behavior as applied to AI evaluation and linguistic QA.

Bachelor’s degree in public relations, advertising, and applied communication completed from 2002 to 2010.