On remote projects for Welo Data, Scale AI, and Prolific, I annotate video and images for computer vision models, including object detection, semantic segmentation, and temporal tracking.
I review frames to maintain consistent object IDs through motion and occlusions, then check for bounding box jitter, missed keyframes, and label inconsistencies before submission.
I also transcribe, translate, and audit audio and video in English, Swahili, and Somali to support NLP model training.
On freelance remote projects, I collected data, evaluated audio, and ranked prompt-response pairs. I followed project guidelines, achieved quality ratings above 98%, and tagged metadata for dataset export.

