AI Output Evaluator — Text, Image & Image Editing
As an AI Output Evaluator on the Starshot, UHRS, and Clickworker platforms, I reviewed and rated AI-generated text and images. Key evaluation criteria included output relevance, coherence, tone, factual accuracy, and instruction alignment. My work supported Reinforcement Learning from Human Feedback (RLHF) pipelines and contributed to improving generative AI performance. • Evaluated AI-generated text for coherence, accuracy, and prompt-following. • Assessed AI-generated images for visual quality and alignment with prompts. • Judged before/after quality in AI image editing and editing consistency across outputs. • Worked with multiple evaluation platforms and content types, adapting to diverse task categories.