Susan Rundell

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Video

Top Label Types

Action Recognition

Freelancer Overview

I am an experienced data annotation and research validation specialist with over four years of hands-on work supporting AI and machine learning projects. My background includes leading large-scale text, image, and video data labeling initiatives and applying structured evaluation rubrics to ensure high-quality, reliable training data. I am skilled in fact-checking, categorizing diverse datasets, and using advanced annotation tools to support domains such as computer vision and natural language processing. My experience collaborating with data science teams has helped me develop a keen eye for detail, reduce error rates, and proactively flag inconsistencies, all while working within fast-paced, high-profile AI training environments. I am proficient in SQL, data reporting, and thrive on delivering accurate, actionable feedback to improve AI model performance.

Labeling Experience

data annotation

CVATImageObject Detection

In my role as a Data Annotation Specialist for an object identification project, I utilized CVAT (Computer Vision Annotation Tool) to transform raw visual data into high-quality training sets for machine learning models. I was responsible for the precise labeling of diverse objects—such as vehicles, pedestrians, and infrastructure—using 2D bounding boxes and complex polygons to ensure pixel-perfect accuracy. By leveraging CVAT’s advanced features, including interpolation for video tracking and attribute tagging for occlusions, I maintained strict data consistency across thousands of frames. My process involved not only manual annotation but also the use of AI-assisted labeling to increase efficiency, followed by a rigorous multi-stage quality assurance review to deliver "ground truth" data with over 98% precision. This technical workflow was critical in reducing model noise and providing the structured information necessary for the AI to navigate and interpret complex real-world enviro

In my role as a Data Annotation Specialist for an object identification project, I utilized CVAT (Computer Vision Annotation Tool) to transform raw visual data into high-quality training sets for machine learning models. I was responsible for the precise labeling of diverse objects—such as vehicles, pedestrians, and infrastructure—using 2D bounding boxes and complex polygons to ensure pixel-perfect accuracy. By leveraging CVAT’s advanced features, including interpolation for video tracking and attribute tagging for occlusions, I maintained strict data consistency across thousands of frames. My process involved not only manual annotation but also the use of AI-assisted labeling to increase efficiency, followed by a rigorous multi-stage quality assurance review to deliver "ground truth" data with over 98% precision. This technical workflow was critical in reducing model noise and providing the structured information necessary for the AI to navigate and interpret complex real-world enviro

2025 - 2025

data annotation

CVATVideoAction Recognition

this project entailed description of vehice behaviour on traffic in asian regions to study mtorists behaviour mostly in asian countries

this project entailed description of vehice behaviour on traffic in asian regions to study mtorists behaviour mostly in asian countries

2022 - 2024

Education

Work History