At Centific, I work on AI and LLM data projects, annotating, labeling, and validating text, image, audio, and video datasets. I review model outputs for relevance, accuracy, and contextual understanding, and provide structured feedback on inconsistencies and quality issues.
I also support prompt-related and language-data evaluation workflows, following project guidelines for accuracy, safety, and consistency.
For my academic project, I developed an audio-based hate speech classifier for short-form videos. I processed 4,746+ videos and compared Logistic Regression, Random Forest, and Support Vector Machines; the optimized Random Forest model achieved up to 78.5% accuracy.

