Dorothy Chepkemoi

Key Skills

Software

Appen

Clickworker

Remotasks

Toloka

Scale AI

Top Subject Matter

Remotask

toloka

appen

Top Data Types

Image

Text

Video

Top Label Types

Classification

Cuboid

Data Collection

Segmentation

Text Generation

Freelancer Overview

I am a seasoned professional with a strong foundation in data labeling and the preparation of AI training datasets. Her keen eye for detail and unwavering commitment to accuracy have been instrumental in improving the precision of machine-learning models across multiple projects. Dorothy's expertise encompasses a broad spectrum of data annotation techniques including semantic segmentation, object classification, and instance identification, ensuring wide-ranging applicability in AI initiatives. My notable contributions include leading a data labeling team in a critical project, where her strategic approach to class imbalance and her innovative methods for enhancing data quality led to a significant increase in model performance. Dorothy's impressive proficiency in utilizing advanced data labeling tools and her ability to work seamlessly with cross-functional AI development teams underline her exceptional qualifications in this field. Her adeptness at quality control, combined with her collaborative nature and results-driven mindset, make her an in-demand expert in the AI data preparation sector.

Labeling Experience

segmentationn

Scale AIVideoBounding BoxPolygon

Frame-by-frame annotation, where individual frames of a video are labeled to provide ground truth data for tasks such as object detection, tracking, and activity recognition. Temporal segmentation, where videos are divided into meaningful segments based on changes in content or activity, allowing for easier analysis and annotation. Semantic segmentation, where objects within the video are classified and labeled in detail, creating pixel-level masks that denote the category of each object. Instance segmentation, a more granular approach that not only classifies objects but also distinguishes between different instances of the same class.

Frame-by-frame annotation, where individual frames of a video are labeled to provide ground truth data for tasks such as object detection, tracking, and activity recognition. Temporal segmentation, where videos are divided into meaningful segments based on changes in content or activity, allowing for easier analysis and annotation. Semantic segmentation, where objects within the video are classified and labeled in detail, creating pixel-level masks that denote the category of each object. Instance segmentation, a more granular approach that not only classifies objects but also distinguishes between different instances of the same class.

2021 - 2023

Labelling

Scale AIImagePolygonPoint Key Point

I played a pivotal role as the lead data annotator, overseeing the intricate process of labeling high volumes of image data for a state-of-the-art AI system. The project aimed to significantly boost the accuracy of an image recognition model used in a critical application, which required precise and comprehensive dataset annotations.

I played a pivotal role as the lead data annotator, overseeing the intricate process of labeling high volumes of image data for a state-of-the-art AI system. The project aimed to significantly boost the accuracy of an image recognition model used in a critical application, which required precise and comprehensive dataset annotations.

2020 - 2023

Education

Work History