Sunday Agbese

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Task Types

Entity (NER) Classification

Freelancer Overview

I am an experienced AI data labeling and evaluation specialist with a strong background in LLM evaluation, prompt rating, safety red teaming, and both text and image annotation across platforms like OneForma, Appen, Clickworker, and Scale AI. My hands-on work includes annotating data for NLP tasks such as NER, span selection, and classification, as well as computer vision projects using tools like CVAT, Label Studio, Labelbox, and Doccano. I am skilled at following detailed rubrics, ensuring policy compliance, and maintaining high quality and accuracy under evolving guidelines and SLAs. Currently pursuing a BSc in Computing with a focus on NLP and Generative AI, I bring a QA mindset, adaptability, and a track record of delivering reliable, high-quality data for AI training and evaluation.

Labeling Experience

Text & Image Annotation for AI Training

CVATVideoEntity Ner Classification

Contributed to multiple AI training projects through platforms such as OneForma, Clickworker, and Prolific, focusing on both NLP and Computer Vision tasks. Text-based work included proofreading, classification, named entity recognition, sentiment analysis, and LLM output evaluation. Image-based projects involved bounding box annotation, object classification, and quality control for e-commerce product imagery. Maintained >95% accuracy rates through adherence to guidelines, double-checking annotations, and incorporating reviewer feedback

Contributed to multiple AI training projects through platforms such as OneForma, Clickworker, and Prolific, focusing on both NLP and Computer Vision tasks. Text-based work included proofreading, classification, named entity recognition, sentiment analysis, and LLM output evaluation. Image-based projects involved bounding box annotation, object classification, and quality control for e-commerce product imagery. Maintained >95% accuracy rates through adherence to guidelines, double-checking annotations, and incorporating reviewer feedback

2023

Education

Work History