AI TUTOR
The scopes of the project were over 5000 AI agent conversations , where I tackled specific data labeling tasks like aligning outputs to original prompts, applying detailed quality to score accuracy, and flagging issues in relevance or safety. It involved granular annotation of text based interactions, constructive feedback loops to tweak model behaviors. The project size constituted of a 5000+ weekly evaluations adding up to tens of thousands monthly.