Mohanprasath Palani

I've completed 2,000+ hours of high-quality text, image, and multimodal annotation for AI training platforms. My work includes RLHF preference ranking and pairwise LLM response comparisons for helpfulness, harmlessness, and honesty.

I annotate text classification, named entities, sentiment, toxicity, images, video, and audio data, maintaining 95%+ accuracy scores and low rejection rates. I work in English and Tamil, bringing cultural sensitivity and localization awareness to language-data projects.

In content moderation projects, I've reviewed user-generated text and images for safety, policy compliance, spam, hate speech, misinformation, adult content, and violence. I've also worked with QA teams to resolve disagreements, document edge cases, and improve labeling guidelines.

My Physics and Law background supports careful reasoning, bias detection, research, and logical consistency checks. I'm available for remote, flexible-hour annotation and AI evaluation work across IST and overlapping US/EU schedules.