AI Content Evaluator (Multimodal)
As an AI Content Evaluator (Multimodal) at Outlier AI, I graded multimodal model outputs according to evolving rubrics, focusing on text-to-image, conversation, and Elo comparison workflows. I maintained a high-quality standard across 5,000+ evaluation tasks and consistently followed project-specific guidelines. My responsibilities emphasized strong written rationales, adaptation to new formats, and precise detection of AI artifacts in generated content. • Conducted evaluation and scoring of AI-generated text and images using project rubrics • Wrote clear, human-toned justification statements for every task • Identified visual and linguistic artifacts specific to AI generation • Adapted to updates in rubrics and surfaced edge cases for rubric improvement.