Freelance AI Trainer & Language Evaluator (Independent / Remote)
Designed and annotated 10,000+ English training samples for RLHF pipelines across multiple LLM client projects. Produced ranked preference data for model comparison tasks to inform reward model training. Authored detailed style guides and quality rubrics to ensure consistency of AI output evaluation. • RLHF preference/ranking dataset creation for enterprise model comparison • Style guide and quality rubric development for annotation teams • Red-teaming exercises and failure-mode reporting to guide safety fine-tuning • Worked with ML engineers to translate linguistic requirements into quantifiable training signals