French Specialist
At Outlier.ai, I contributed to AI training and evaluation projects, primarily as a French language specialist. My work involved: Prompt creation – Designing clear, targeted prompts in both French and English to elicit high-quality responses from AI models. Response evaluation – Reviewing and scoring AI outputs for grammatical accuracy, factual correctness, clarity, and relevance to the prompt. Conversational training – Simulating natural, human-like dialogues to help the AI improve flow, tone, and contextual awareness. Localization – Adapting AI responses for cultural and linguistic appropriateness, ensuring they align with French-speaking audiences. AI safety and red-teaming – Creating challenging or edge-case scenarios to test AI safety mechanisms and refine content moderation filters. Annotation and categorization – Labeling datasets, classifying text, and adding metadata to improve AI comprehension and retrieval accuracy. Error analysis – Identifying recurring AI weaknesses