AI Data Trainer & AnnotatorFreelance / Remote (Argentina)Project Objective: Improved the accuracy, helpfulness, and safe
AI Data Trainer & AnnotatorFreelance / Remote (Argentina)Project Objective: Improved the accuracy, helpfulness, and safety of a leading Large Language Model (LLM) through Reinforcement Learning from Human Feedback (RLHF).Data Labeling & Annotation: Categorized and labeled complex textual datasets, focusing on semantic intent, factual accuracy, and bias detection.Model Evaluation: Reviewed AI-generated responses, ranked multiple outputs based on strict guidelines, and rewrote answers to train the model on nuanced human conversation.Prompt Engineering: Crafted diverse and challenging prompts to stress-test the AI's reasoning capabilities, coding logic, and contextual understanding.Quality Assurance: Maintained a consistent accuracy rating above 95% by applying rigorous data taxonomy rules and identifying edge cases.