Duncan Kahiga

Key Skills

Software

Appen

Clickworker

CloudFactory

Data Annotation Tech

Other

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Label Types

No label types listed

Freelancer Overview

I have extensive experience creating and refining high-quality datasets for AI/ML model training across multiple domains, including natural language processing and computer vision. My core competency lies in executing complex annotation tasks—such as semantic segmentation, bounding boxes, entity recognition, and sentiment labeling—with a consistent emphasis on accuracy and detail. I am proficient in using leading labeling platforms (e.g., Labelbox, CVAT, and proprietary tools) and have a strong understanding of labeling guidelines, edge cases, and quality assurance protocols to ensure data integrity. What sets me apart is my ability to bridge raw data processing with model performance insight. For example, in a key project for a document understanding AI, I not only managed a team to label thousands of legal documents for key entities, but I also analyzed recurring discrepancies in the annotations, leading to a revision of the guidelines that improved inter-annotator agreement by 25% and directly enhanced the model's F1-score. This combination of hands-on precision, project oversight, and a analytical, impact-driven approach allows me to contribute directly to the development of more robust and effective AI systems.

Labeling Experience

Human movement

AppenVideoBounding BoxPoint Key Point

Vast movement and mapping.

Vast movement and mapping.

2025 - 2025

Education

Work History