Rufus Obet

Key Skills

Software

CloudFactory

CVAT

Data Annotation Tech

Labelbox

Mercor

Remotasks

Scale AI

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Video

Top Label Types

Bounding Box

Polygon

Point Key Point

Segmentation

Freelancer Overview

I am an applied AI engineer with a strong foundation in mathematics and over five years of experience building, evaluating, and deploying machine learning solutions. My work has focused on the creation and curation of high-quality training datasets, including synthetic and augmented data generation to address data sparsity and improve model performance. I have hands-on experience with data labeling, annotation, and evaluation using Python, pandas, NumPy, and specialized ML tools such as scikit-learn, PyTorch, and Hugging Face Transformers. In several projects, I have built retrieval-augmented generation (RAG) systems and designed robust evaluation workflows to ensure data quality and model reliability. My background also includes translating business requirements into clear data annotation guidelines and delivering workshops to both technical and non-technical audiences, ensuring that data-centric AI solutions meet real-world needs. I am passionate about creating data pipelines and evaluation frameworks that drive trustworthy AI outcomes across domains like analytics, document understanding, and knowledge retrieval.

Labeling Experience

Data Associate

CVATImageBounding BoxPolygon

This project involved labeling images for machine learning model training, focusing on vehicle, human, and large language model (LLM) related datasets. Tasks included annotating images using bounding boxes and point/key points to precisely identify and mark key features. The project followed quality standards to ensure accurate and consistent labeling, adhering to defined labeling guidelines. The dataset size ranged from thousands to tens of thousands of images, with regular quality checks and validation processes in place to maintain high labeling accuracy. The data labeling was performed using the CVAT tool, ensuring efficient and streamlined workflows.

This project involved labeling images for machine learning model training, focusing on vehicle, human, and large language model (LLM) related datasets. Tasks included annotating images using bounding boxes and point/key points to precisely identify and mark key features. The project followed quality standards to ensure accurate and consistent labeling, adhering to defined labeling guidelines. The dataset size ranged from thousands to tens of thousands of images, with regular quality checks and validation processes in place to maintain high labeling accuracy. The data labeling was performed using the CVAT tool, ensuring efficient and streamlined workflows.

2019 - 2025

Education

Work History