Project: Rufus – Al Shopping Assistant LLM and Prompt Evaluation Evaluate Rufus LLM responses against defined quality criteria (accuracy, relevance, completeness, consistency, helpfulness, intent alignment) and give structured feedback that drives model improvement. Analyze customer prompts and prompt-response pairs to flag ambiguous prompts, misread intent, irrelevant answers, factual inconsistencies, and contextual mismatches. Assess how customer queries, product information, and search/recommendation results relate to Al responses, pinpointing gaps in response relevancy. Create ground-truth data and perform annotation, curation, classification, and labeling for shopping-related AI/ML datasets. Monitor Al watchlist and quality workflows, flagging problematic outputs and enforcing evaluation standards across high-volume data operations. Data Quality, Analytics and Reporting Executed ETL processes to extract, transform, load, and validate Al model and business datasets, improving data reliability and boosting response quality by 75%. Perform Root Cause Analysis (RCA) with SQL and Excel to trace data issues, identify recurring patterns, and resolve quality problems. Run data quality assessments, auditing, validation, and UAT across AI/ML workflows; document defects and verify fixes through structured testing. Manage large-scale multimodal datasets (text, speech, image, video) for Alexa and Rufus, using Atlas and Mercury to clean, structure, and track annotation workflows. Build and maintain Power BI dashboards that turn raw data into KPI visualizations and actionable insights for stakeholders. Process and Collaboration Write and maintain SOPs, process maps, and quality guidelines that improve workflow consistency, scalability, and compliance with PMO and data governance standards. Partner with Al, data, operations, and engineering teams to surface quality issues, investigate root causes, and clarify requirements; track defects and resolutions in JIRA and Slack. Awards and Achievements High Flyer Award, Amazon Development Centre: recognized for outstanding performance and measurable impact on ML data quality and operational efficiency