At Teleperformance, I audit data labeled by human reviewers to train AI decision-making, evaluate LLM outputs, and prepare AI accuracy reports. My RCA-driven trainings, guideline clarifications, and operational changes helped raise AI accuracy from 85% to 96%.
I audit 150 reviewers per week as part of a six-person QA team. I also run calibration sessions, track results in Excel and Google Sheets, and coach reviewers in English.
Previously, at Teleperformance, I reviewed online content against predefined policies and contributed to policy improvement. At Infomag Publishing, I led digital projects for Harvard Business Review Türkiye, Bloomberg Businessweek Türkiye, and CoinDesk Türkiye, with performance analysis and reporting.

