Hire this AI Trainer
Sign in or create an account to invite AI Trainers to your job.
Labelbox
CVATNo subject matter listed
Audio
Computer Code ProgrammingOver the last three years I have produced, labelled and quality-controlled 120 k+ images and 2 k+ hours of video for computer-vision and NLP pipelines spanning autonomous driving, retail analytics and document-AI. My live Labelbox dashboard (37 reviews, 3.6 star average, 5.0 ⭐ on red-teaming and audio) shows I deliver pixel-perfect labels while meeting tight deadlines. Daily power-user of CVAT, Labelbox, Prodigy and FiftyOne, I write lightweight Python utilities (OpenCV, Pandas, Pillow) that cut annotation time 30 % and enforce 99 % inter-annotator agreement. I understand how label noise propagates to mAP, so I script consistency checks, auto-pre-labeling and version-controlled exports (COCO, YOLO, Pascal-VOC) that plug straight into S3/GCP training pipelines.
4-reviewed batch, French audio clips, toxicity score + target-group flag. Perfect 5.0 Labelbox rating (zero revisions). Reduced client moderation false-negative rate by 22 %.
5-reviewed batch of French video-text alignment (OCR + subtitle timing). Achieved 3.8 (above platform median) with 0.7 % revision rate. Edge-cases (slang, ambient noise) flagged and re-submitted for next cycle.
3 k French audio clips, toxicity + target-group flags. Achieved perfect 5.0 Labelbox rating (4 reviews, Sep 2025). Reduced client moderation false-negative rate by 22 %.
4 k infrared frames, instance masks for pedestrians + attribute “carrying bag”. Perfect 5.0 Labelbox rating (4 reviews). Reduced client false-positive rate by 18 %.
Annotated 35 k highway images with 80 k instances (cars, trucks, buses, lanes, road signs). Tasks: 2-D bounding boxes, lane polygons, truncation/occlusion flags. Achieved 99.1 % consensus, 0.7 % revision rate. Dataset improved client YOLOv8 mAP by +4.2 pts.
Bachelor of Science, Probability and Vector Spaces
N/A, Information Technology Development
Front-End/Back-End Developer
Sales Representative
