Joan Asekenye

Key Skills

Software

CVAT

Labelbox

Sama

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Label Types

Bounding Box

Polygon

Cuboid

Freelancer Overview

I am an experienced IT specialist with over 8 years working in data entry, online research, and customer support roles across technology, energy, and healthcare sectors. My background includes hands-on data management, updating and maintaining client Excel sheets, and ensuring high-quality, accurate data capture for business operations. I am skilled in using computer systems, applications, and KYC guidelines to monitor and maintain data integrity, and I have a strong commitment to detail and client satisfaction. My experience collaborating with teams via email and Slack, along with my ability to troubleshoot and resolve issues efficiently, makes me well-suited for data labeling and AI training data roles. I am fluent in multiple languages and bring strong communication, problem-solving, and organizational skills to every project I undertake.

Labeling Experience

Data annotator

LabelboxVideoBounding BoxPolygon

The road-sign-detection (RSD) project aimed to analyze video data for the automotive sector, distinguishing over 150 road signs/traffic lights and 15 physical objects using a dataset of approximately 35,000 images and 50,000 labels. Data labeling involved frame extraction, manual bounding box annotation, numerical labeling, and augmentation techniques like rotation and zooming. Quality measures prioritized detection accuracy using the Faster R-CNN architecture, with regular dataset analysis, automated image validation, and visualization of class frequencies contributing to an approximately 97% accuracy for road sign detection at close proximity

The road-sign-detection (RSD) project aimed to analyze video data for the automotive sector, distinguishing over 150 road signs/traffic lights and 15 physical objects using a dataset of approximately 35,000 images and 50,000 labels. Data labeling involved frame extraction, manual bounding box annotation, numerical labeling, and augmentation techniques like rotation and zooming. Quality measures prioritized detection accuracy using the Faster R-CNN architecture, with regular dataset analysis, automated image validation, and visualization of class frequencies contributing to an approximately 97% accuracy for road sign detection at close proximity

2020 - 2023

Education

Work History