Rebecca Obongo

Key Skills

Software

CloudFactory

CVAT

Data Annotation Tech

Dataloop

Labelbox

LabelImg

Mindrift

OneForma

Remotasks

Sama

Scale AI

Toloka

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Classification

Point/Key Point

Polygon

Polyline

Freelancer Overview

I have hands-on experience in data labeling and AI training support, with a strong background in tasks involving text classification, search relevance evaluation, and content accuracy review. My work has focused on ensuring high-quality, consistent, and contextually accurate annotations, particularly for English-language datasets. I am skilled in following detailed guidelines, maintaining accuracy under tight deadlines, and applying strong analytical thinking to improve model performance. In addition to annotation work, I bring experience in quality assurance, research, documentation, and digital operations. I have developed strong attention to detail, excellent communication skills, and the ability to work both independently and collaboratively. My familiarity with Kenyan cultural, linguistic, and online search patterns enables me to provide highly relevant and localized annotations for AI systems. I am committed to producing clean, reliable data that contributes to the development of smarter, safer, and more accurate AI models.

Labeling Experience

Video Annotation

Scale AIVideoBounding Box

The project involved supporting the development of AI systems by providing high-quality labeled data across text, image, and search-relevance tasks. My responsibilities included annotating large datasets, reviewing and validating data for accuracy, classifying content, evaluating search intent, and ensuring relevance according to detailed project guidelines. The project size ranged from thousands to tens of thousands of data points, requiring consistent attention to detail and efficient workflow management. To maintain quality, I followed strict annotation protocols, adhered to QA benchmarks, and completed regular accuracy checks. I also incorporated feedback from reviewers to improve consistency and achieve target quality thresholds, typically above 95% accuracy. My work contributed to creating reliable training data that improved model performance and alignment with user expectations.

The project involved supporting the development of AI systems by providing high-quality labeled data across text, image, and search-relevance tasks. My responsibilities included annotating large datasets, reviewing and validating data for accuracy, classifying content, evaluating search intent, and ensuring relevance according to detailed project guidelines. The project size ranged from thousands to tens of thousands of data points, requiring consistent attention to detail and efficient workflow management. To maintain quality, I followed strict annotation protocols, adhered to QA benchmarks, and completed regular accuracy checks. I also incorporated feedback from reviewers to improve consistency and achieve target quality thresholds, typically above 95% accuracy. My work contributed to creating reliable training data that improved model performance and alignment with user expectations.

2018 - 2024

Agent

Scale AIImageSegmentation

The project involved large-scale image segmentation to support computer vision model development for an AI client. My main tasks included drawing precise pixel-level polygons, annotating object boundaries, identifying multiple classes within complex scenes, and verifying segment masks for accuracy. I worked on a high-volume dataset containing thousands of images across several batches, each requiring strict adherence to detailed labeling guidelines. To maintain quality, I consistently met the project’s required accuracy benchmarks, performed self-QA checks, corrected inconsistencies, and incorporated feedback from the quality team. This ensured that all segmentation outputs were clean, consistent, and reliable for model training and validation.

The project involved large-scale image segmentation to support computer vision model development for an AI client. My main tasks included drawing precise pixel-level polygons, annotating object boundaries, identifying multiple classes within complex scenes, and verifying segment masks for accuracy. I worked on a high-volume dataset containing thousands of images across several batches, each requiring strict adherence to detailed labeling guidelines. To maintain quality, I consistently met the project’s required accuracy benchmarks, performed self-QA checks, corrected inconsistencies, and incorporated feedback from the quality team. This ensured that all segmentation outputs were clean, consistent, and reliable for model training and validation.

2018 - 2019

Image Anntation

SamaImageBounding Box

The project involved supporting the development of AI systems by providing high-quality labeled data across text, image, and search-relevance tasks. My responsibilities included annotating large datasets, reviewing and validating data for accuracy, classifying content, evaluating search intent, and ensuring relevance according to detailed project guidelines. The project size ranged from thousands to tens of thousands of data points, requiring consistent attention to detail and efficient workflow management. To maintain quality, I followed strict annotation protocols, adhered to QA benchmarks, and completed regular accuracy checks. I also incorporated feedback from reviewers to improve consistency and achieve target quality thresholds, typically above 95% accuracy. My work contributed to creating reliable training data that improved model performance and alignment with user expectations.

The project involved supporting the development of AI systems by providing high-quality labeled data across text, image, and search-relevance tasks. My responsibilities included annotating large datasets, reviewing and validating data for accuracy, classifying content, evaluating search intent, and ensuring relevance according to detailed project guidelines. The project size ranged from thousands to tens of thousands of data points, requiring consistent attention to detail and efficient workflow management. To maintain quality, I followed strict annotation protocols, adhered to QA benchmarks, and completed regular accuracy checks. I also incorporated feedback from reviewers to improve consistency and achieve target quality thresholds, typically above 95% accuracy. My work contributed to creating reliable training data that improved model performance and alignment with user expectations.

2018 - 2019

Education

Work History