Gedion Kiprop

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Top Label Types

Polygon

Object Detection

Freelancer Overview

I am a detail-oriented data annotator with strong experience in image labeling and polygon annotation for computer vision projects. My expertise includes accurately identifying UI buttons, icons, and objects in images, even under challenging angles and lighting conditions. I consistently maintain over 98% labeling accuracy and can handle large datasets, completing up to 250 images per day and over 3,000 images per project. I am proficient with annotation tools such as CVAT, LabelImg, Roboflow, Supervisely, Labelbox, and VGG VIA, and have a solid background in data QA and dataset management. I am committed to delivering high-quality, precise annotations on tight deadlines and am available immediately for new projects.

Labeling Experience

UI Button & Object Polygon Annotation for Computer Vision Model

CVATImagePolygonObject Detection

Worked on a computer vision dataset focused on annotating UI buttons, icons, and interface objects using precise polygon annotation techniques. The project required labeling 5–6 objects per image, including both icon shapes and associated text elements under various angles and inclined perspectives. Scope of work included: Accurate polygon tracing of buttons and objects Identification and separation of icon and text components Multi-object labeling per image Quality control review before submission Following strict annotation guidelines to ensure dataset consistency Project size involved annotating over 2,500+ objects across thousands of images. Maintained high accuracy standards (97%+ internal quality checks) and ensured clean, consistent annotations to improve AI object detection performance. Focused on balancing speed and precision to meet tight deadlines while maintaining dataset integrity for model training.

Worked on a computer vision dataset focused on annotating UI buttons, icons, and interface objects using precise polygon annotation techniques. The project required labeling 5–6 objects per image, including both icon shapes and associated text elements under various angles and inclined perspectives. Scope of work included: Accurate polygon tracing of buttons and objects Identification and separation of icon and text components Multi-object labeling per image Quality control review before submission Following strict annotation guidelines to ensure dataset consistency Project size involved annotating over 2,500+ objects across thousands of images. Maintained high accuracy standards (97%+ internal quality checks) and ensured clean, consistent annotations to improve AI object detection performance. Focused on balancing speed and precision to meet tight deadlines while maintaining dataset integrity for model training.

2025

Education

Work History