Hans Juma

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Top Label Types

Bounding Box

Polygon

Freelancer Overview

I am a multilingual AI data specialist with over 5 years of hands-on experience in data annotation, labeling, transcription, and audio editing for English and Swahili projects. My background includes preparing high-quality multilingual datasets for AI model training, supporting LLM fine-tuning (using no-code platforms), and collaborating with global teams to deliver accurate, culturally relevant results under tight deadlines. I have worked on diverse projects involving text, audio, and image data, with a strong focus on natural language processing and speech recognition for African languages. My expertise extends to using tools like Zoom, Slack, Trello, and Google Workspace for remote collaboration, and I am committed to upholding data quality and ethical standards in every project.

Labeling Experience

Data annotator

CVATImageBounding BoxPolygon

I worked on a data labeling project for autonomous vehicles where I annotated street footage, drawing boxes around cars, pedestrians, traffic lights, and road signs to help the vision models learn how to interpret real-world driving scenes. At the same time, I also supported an LLM project where I reviewed and labeled text data, rated responses for quality and accuracy, and flagged edge cases to improve model performance. It was detailed, quality-focused work, but really interesting knowing the data was directly shaping how both AI systems see and understand the world.

I worked on a data labeling project for autonomous vehicles where I annotated street footage, drawing boxes around cars, pedestrians, traffic lights, and road signs to help the vision models learn how to interpret real-world driving scenes. At the same time, I also supported an LLM project where I reviewed and labeled text data, rated responses for quality and accuracy, and flagged edge cases to improve model performance. It was detailed, quality-focused work, but really interesting knowing the data was directly shaping how both AI systems see and understand the world.

2025 - 2025

Education

Work History