Kennedy Lunani

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Video

Top Label Types

Bounding Box

Freelancer Overview

I am a detail-oriented Video Annotation Specialist with hands-on experience labeling and tagging video data for AI and computer vision projects. My expertise includes object detection, tracking, segmentation, and classification using industry-standard tools such as Labelbox, CVAT, VGG, and Supervisely. I have successfully annotated thousands of video frames for machine learning datasets, ensuring high accuracy and consistency through rigorous quality checks. I am committed to delivering reliable, high-quality datasets on time and thrive in remote, independent work environments. My strong attention to detail and adaptability make me an asset to any data labeling or AI training data team.

Labeling Experience

OBJECT DETECTION

CVATVideoBounding Box

Project Scope The project focused on preparing high-quality labeled datasets for computer vision models, mainly for object detection applications. The objective was to accurately identify and localize target objects within images through bounding box annotation, supporting the development of reliable AI systems for real-world deployment. The dataset included a wide range of environments, lighting conditions, object orientations, and background complexity to ensure strong model generalization. Specific Data Labeling Tasks Performed The annotation work involved drawing tight and accurate bounding boxes around predefined object categories and labeling multiple objects per image while maintaining class consistency. Special attention was given to handling occluded, truncated, and overlapping objects by placing boundaries correctly according to project rules. The work followed detailed labeling guidelines such as partial visibility instructions, minimum object size thresholds, and standar

Project Scope The project focused on preparing high-quality labeled datasets for computer vision models, mainly for object detection applications. The objective was to accurately identify and localize target objects within images through bounding box annotation, supporting the development of reliable AI systems for real-world deployment. The dataset included a wide range of environments, lighting conditions, object orientations, and background complexity to ensure strong model generalization. Specific Data Labeling Tasks Performed The annotation work involved drawing tight and accurate bounding boxes around predefined object categories and labeling multiple objects per image while maintaining class consistency. Special attention was given to handling occluded, truncated, and overlapping objects by placing boundaries correctly according to project rules. The work followed detailed labeling guidelines such as partial visibility instructions, minimum object size thresholds, and standar

2023 - 2024

Education

Work History