AI Quality Analyst
Scope of Project: Evaluated personalized AI interactions for Gemini, testing how well it used user data from past chats, Gmail, Search, and YouTube to provide relevant responses. Quality Measures: Assessed responses for grounding, integration, and helpfulness; performed side-by-side comparisons and wrote clear rationales; maintained strict data hygiene. Project Size: Handled hundreds of multi-turn conversational evaluations weekly as part of a cross-functional team. Data Labeling Tasks: Designed prompts, annotated model responses, ranked outputs SxS, verified debug info, and deleted evaluation data to ensure accuracy and privacy.