Natural Response Annotation for Conversational AI (French/English)
Worked on a large-scale AI training project focused on enhancing the performance of multilingual conversational models (French and English). The scope involved natural response generation, annotation, and correction to feed supervised learning systems. Key tasks included: Creating and labeling natural, human-like responses for diverse prompts Performing intent classification, sentiment tagging, and coherence validation Reviewing AI-generated outputs for fluency, accuracy, and tone Ensuring outputs adhered to strict project-specific linguistic guidelines Collaborating remotely with QA leads to meet weekly accuracy targets (>95%)