AI Content Evaluator & Data Annotator (Freelance)
Remotasks
Text
RLHFServed as an AI content evaluator and data annotator for RLHF training workflows by reviewing and ranking AI-generated responses. Labeled datasets with semantic and behavioral signals such as intent and sentiment, and provided safety and quality judgments to improve model performance. Conducted preference-based evaluations to guide reward model training and reduced policy-violating content in model outputs. • Evaluated responses for quality, accuracy, coherence, and safety. • Applied semantic labels, intent classification, and sentiment scoring for fine-tuning. • Performed preference rankings from side-by-side comparisons for reward model training. • Flagged harmful, biased, or factually incorrect outputs and documented rationale.
2022 - Present