Dequan Johnson

Key Skills

Software

Appen

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Top Label Types

Text Summarization

Diagnosis

Bounding Box

Classification

Evaluation Rating

Freelancer Overview

I am an experienced AI evaluation specialist and electrical engineer with a strong background in designing and validating high-quality training data for generative AI systems. Over the past three years, I have created and assessed 200+ complex technical prompts and evaluated more than 1,000 multi-step AI-generated solutions using objective rubrics, deterministic scoring, and Python-based validation tools such as NumPy, SciPy, and Pandas. My work has focused on engineering and technical domains, where I have improved reviewer agreement and reduced turnaround times by developing reproducible QA workflows and structured grading systems. I am skilled in prompt design, multi-step reasoning validation, and the creation of robust annotation guidelines, and I bring a detail-oriented approach to ensuring data accuracy and consistency in AI training projects.

Labeling Experience

Data Labeling Contributor

AppenImageText SummarizationDiagnosis

Contributed to a medium-scale AI training project involving over 50,000 text entries sourced from web searches and chatbot interactions. Labeled each entry for sentiment (positive, negative, neutral), topic category, and relevance to user intent. Ensured strict adherence to Appen’s quality guidelines, maintaining high labeling accuracy and consistency through regular review cycles and peer comparisons. Monitored labeling quality metrics and flagged ambiguous or unclear entries for resolution, contributing to the creation of a high-quality, structured dataset for supervised machine learning and natural language processing models.

Contributed to a medium-scale AI training project involving over 50,000 text entries sourced from web searches and chatbot interactions. Labeled each entry for sentiment (positive, negative, neutral), topic category, and relevance to user intent. Ensured strict adherence to Appen’s quality guidelines, maintaining high labeling accuracy and consistency through regular review cycles and peer comparisons. Monitored labeling quality metrics and flagged ambiguous or unclear entries for resolution, contributing to the creation of a high-quality, structured dataset for supervised machine learning and natural language processing models.

2024 - 2024

Data Annotation Specialist

TelusImageBounding BoxClassification

Contributed to a medium-scale annotation project involving over 40,000 images and text entries for AI model training. Annotated images with bounding boxes, segmentation, and object tagging, and labeled text data for content classification and relevance evaluation. Followed TELUS AI’s quality standards to ensure accurate, consistent labeling and flagged ambiguous entries for further review. Produced high-quality, structured datasets used for supervised machine learning in computer vision and natural language processing applications.

Contributed to a medium-scale annotation project involving over 40,000 images and text entries for AI model training. Annotated images with bounding boxes, segmentation, and object tagging, and labeled text data for content classification and relevance evaluation. Followed TELUS AI’s quality standards to ensure accurate, consistent labeling and flagged ambiguous entries for further review. Produced high-quality, structured datasets used for supervised machine learning in computer vision and natural language processing applications.

2023 - 2024

Education

Work History