Phương Thu

Detail-oriented AI Model Evaluation and Quality Evaluation professional with 4+ years of experience in content evaluation, quality assurance, policy-based decision-making, and Trust & Safety operations. Experienced in evaluating AI model outputs using audio and text data, conducting side-by-side model comparisons, assessing factual and contextual accuracy, and identifying inconsistencies between original inputs and AI-generated outputs. Proven ability to apply detailed evaluation guidelines consistently while maintaining 98%+ quality accuracy in high-volume environments.

Strong analytical thinking, attention to detail, written communication, and experience working independently in remote, guideline driven projects.

CONTENT MODERATOR at Gear Inc (2021-12 – 2026-03)

• Evaluated text, images, videos, and user-generated content according to platform policies and Community Standards.
• Applied detailed policy guidelines to make accurate and consistent decisions in nuanced and ambiguous cases.
• Identified and escalated policy violations involving misinformation, fraud, hate speech, violence, and adult content.
• Maintained 98%+ quality accuracy while consistently meeting high-volume productivity targets.
• Reviewed approximately 800–1,200 pieces of content per shift while maintaining evaluation quality and attention to detail.
• Collaborated with QA teams to improve evaluation consistency, quality, and guideline adherence.
• Demonstrated strong analytical thinking and decision-making when evaluating content under time constraints.
• Consistently exceeded 110% productivity targets and achieved Top Performer recognition for multiple consecutive months.

AI Model Evaluation - Audio & Text at CONFIDENTIAL SHORT-TERM PROJECT

• Conducted multi-turn interactions with AI models to evaluate response quality, factual accuracy, and contextual consistency.
• Performed side-by-side evaluations of different AI model outputs and assessed comparative response quality based on project-specific evaluation criteria.
• Compared AI-generated outputs with original audio inputs to identify discrepancies, inaccuracies, inconsistencies, and alignment issues.
• Evaluated audio and text data with strong attention to detail and consistency.
• Interacted directly with AI models to provide corrections and contextual information to support the evaluation process.
• Applied detailed project-specific guidelines consistently throughout the evaluation process.

Bachelor in Hospitality Management – Dong A University (2017 – 2021)