Data Labeling Specialist (AI Data Trainer) at DataAnnotation, Remote
Text
RLHFServed as an AI Data Trainer conducting large-scale AI training, evaluation, and language model optimization tasks through NLP annotation workflows. Evaluated and refined AI-generated responses using Reinforcement Learning from Human Feedback (RLHF) methodologies to improve accuracy, alignment, and safety compliance. Authored and tested advanced technical prompts and complex Python/C++ coding tasks, performing QA on logical reasoning, fact consistency, and multi-turn conversation quality. • Evaluated logical and programming capabilities of neural networks using code/logic-focused prompts. • Performed fact-checking and logical flow analysis to reduce hallucinations. • Designed edge-case scenarios and stress-tested models via adversarial prompting to find vulnerabilities. • Conducted multi-turn conversation evaluations for robustness and safety alignment.