NELSON SONGOK

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Entity Ner Classification

Freelancer Overview

With over 7 years of IT experience, I have developed strong expertise in data management, technical support, and digital documentation, which directly translates to high-quality data labeling and annotation work. My recent roles involved capturing and organizing 360° visual data using OpenSpace, Insta360, and Ricoh Theta cameras, as well as managing and validating site documentation for construction and mechanical projects through platforms like Procore, Bluebeam Revu, Navisworks, and Autodesk BIM 360. I am skilled in reviewing and annotating complex visual datasets, ensuring accuracy for digital twins and BIM models, and maintaining meticulous records for future reference. My background in network administration, cybersecurity, and technical troubleshooting gives me a keen eye for detail and data integrity, making me well-suited for AI training data and annotation projects, especially in computer vision and structured data environments. I am committed to delivering precise, reliable data to support the development of robust AI and machine learning solutions.

Labeling Experience

Data Labeling & AI Training Specialist

CVATTextEntity Ner Classification

The project focused on improving NLP model performance through high-quality text annotation. My role involved Named Entity Recognition (NER), intent classification, sentiment labeling, and content moderation tagging on conversational and short-form text data. I annotated and reviewed several thousand text records following detailed project guidelines and labeling taxonomies. Quality measures included double-pass reviews, inter-annotator agreement checks, adherence to strict annotation guidelines, and regular accuracy audits to ensure consistency, precision, and policy compliance

The project focused on improving NLP model performance through high-quality text annotation. My role involved Named Entity Recognition (NER), intent classification, sentiment labeling, and content moderation tagging on conversational and short-form text data. I annotated and reviewed several thousand text records following detailed project guidelines and labeling taxonomies. Quality measures included double-pass reviews, inter-annotator agreement checks, adherence to strict annotation guidelines, and regular accuracy audits to ensure consistency, precision, and policy compliance

2025 - 2025

Education

Work History