Image Evaluation & Rating
Task: Rated 1,000+ images for relevance, clarity, and safety based on strict client guidelines (e.g., content moderation, object detection, or facial recognition training). Process: Labeled images using tools like Labelbox, V7, or custom annotation platforms, tagging objects, identifying unsafe content, or assigning quality scores (e.g., 1–5 scale). Outcome: Achieved 99% inter-rater agreement and zero guideline violations, improving model accuracy for an AI computer vision project. Key skill applied: Following detailed labelling rubrics, maintaining consistency across large batches, and meeting daily annotation targets.