Roy Mwikya

Key Skills

Software

Appen

Axiom AI

Clickworker

CloudFactory

CVAT

Img Lab

LabelImg

V7 Labs

Labelbox

Top Subject Matter

Computer Vision: Experience in labeling data for self-driving cars, facial recognition, and object detection in retail.

Natural Language Processing (NLP): Expertise in sentiment analysis, chatbots, and large language model evaluation.

Healthcare AI: Annotated medical images (X-rays, MRIs) and patient data for diagnostic tools.

Top Data Types

Audio

Geospatial Tiled Imagery

Image

Top Task Types

Bounding Box

Cuboid

Data Collection

Fine Tuning

Freelancer Overview

I am an experienced data labeling specialist with over five years in AI training data annotation, particularly in computer vision and natural language processing (NLP). Skilled in annotating complex image and text data, I have contributed to high-impact projects, such as enhancing self-driving car algorithms through precise object detection, training facial recognition models, and developing chatbots using sentiment analysis. My versatility extends to the healthcare sector, where I’ve annotated medical images, aiding in diagnostic AI solutions that support improved patient outcomes. With a strong foundation in quality control, I ensure each annotation meets accuracy standards essential for robust AI model performance. I am proficient in using industry-standard annotation tools and optimizing workflows to boost labeling efficiency. My commitment to detail and ability to adapt to diverse project needs make me a valuable asset in any AI training data project.

Labeling Experience

Medical Imaging for Diagnostic AI

V7 LabsVideoClassificationObject Detection

Annotated medical imaging data, including X-rays and MRIs, to assist in the training of diagnostic AI models. Applied detailed segmentation and tagging for various conditions, enhancing the model's accuracy in detecting anomalies like fractures, tumors, and infections, which improved clinical decision-making tools for healthcare providers.

Annotated medical imaging data, including X-rays and MRIs, to assist in the training of diagnostic AI models. Applied detailed segmentation and tagging for various conditions, enhancing the model's accuracy in detecting anomalies like fractures, tumors, and infections, which improved clinical decision-making tools for healthcare providers.

2024

NLP and computer vision models

LabelboxComputer Code Programming

Designed and implemented training workflows for NLP and computer vision models, achieving a 15% improvement in model accuracy across projects. • • Curated, annotated, and validated large-scale datasets using tools like Labelbox and Prodigy. • Collaborated with engineers to refine machine learning models, ensuring optimal performance in production environments. • Conducted quality assurance checks on model outputs, identifying errors and providing actionable feedback to improve accuracy. • Developed internal guidelines for ethical AI training practices and bias mitigation strategies.

Designed and implemented training workflows for NLP and computer vision models, achieving a 15% improvement in model accuracy across projects. • • Curated, annotated, and validated large-scale datasets using tools like Labelbox and Prodigy. • Collaborated with engineers to refine machine learning models, ensuring optimal performance in production environments. • Conducted quality assurance checks on model outputs, identifying errors and providing actionable feedback to improve accuracy. • Developed internal guidelines for ethical AI training practices and bias mitigation strategies.

2024 - 2024

Autonomous Driving Data Labeling

Axiom AIVideoObject DetectionAction Recognition

Worked on a large-scale project for self-driving car AI, annotating video frames to identify pedestrians, vehicles, road signs, and lane markings. Used bounding boxes, semantic segmentation, and keypoint annotations to support real-time decision-making capabilities in autonomous vehicle systems.

Worked on a large-scale project for self-driving car AI, annotating video frames to identify pedestrians, vehicles, road signs, and lane markings. Used bounding boxes, semantic segmentation, and keypoint annotations to support real-time decision-making capabilities in autonomous vehicle systems.

2023 - 2024

Education

Work History