IT Engineer & AI Data Specialist at Toontechs Services (Data labeling and AI evaluation)
Created and enforced a validation framework to improve structured data annotation accuracy across 10,000+ monthly records, reducing labeling errors. Applied systematic reasoning analysis to evaluate and annotate 500+ structured AI outputs per week, including hallucination, logic-gap, and off-task detection before delivery. Designed prompt templates and instruction-following rubrics to assess and correct AI-generated responses across 200+ daily tasks. • Validated structured outputs against quality benchmarks and QA criteria. • Measured performance via internal QA benchmarks and weekly evaluation scores. • Conducted near-zero error rate quality assurance through repeatable evaluation workflows. • Supported iterative improvement of labeling and evaluation processes for client delivery.