Joseph Kilonzo

Key Skills

Software

Appen

Clickworker

Lionbridge

OneForma

Remotasks

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Video

Top Task Types

Action Recognition

Bounding Box

Data Collection

Evaluation Rating

Object Detection

Freelancer Overview

Since 2021, I have contributed to a wide range of data labeling and training projects on platforms such as Outlier, Appen, OneForma, Prolific, and Remotasks. My experience spans language-focused work (including projects like Nightingale), as well as subject-specific annotation in History, Biology, and Chemistry. I have also carried out extensive academic-style research tasks on Prolific, strengthening my ability to work with complex and specialized content. Alongside text and research projects, I have supported image, video, and 3D/LiDAR annotation for self-driving and vision-based systems. I am detail-oriented, dependable, and quick to learn new guidelines. With fluency in English and Kiswahili, plus practical skills in Excel and Kobo Toolbox, I bring both accuracy and cultural understanding to diverse projects while continuing to grow my expertise across data domains.

Labeling Experience

LLM Evaluation and Text Annotation

Scale AITextQuestion AnsweringText Summarization

Contributed to Project Nightingale, a large-scale language model evaluation and annotation initiative. Tasks included ranking AI-generated responses, identifying factual accuracy, improving fluency and coherence, and labeling text for sentiment and named entities. The project also involved prompt and response writing to support supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). Ensured consistency and quality by adhering to platform-specific guidelines, achieving accuracy scores above 95% across multiple review cycles.

Contributed to Project Nightingale, a large-scale language model evaluation and annotation initiative. Tasks included ranking AI-generated responses, identifying factual accuracy, improving fluency and coherence, and labeling text for sentiment and named entities. The project also involved prompt and response writing to support supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). Ensured consistency and quality by adhering to platform-specific guidelines, achieving accuracy scores above 95% across multiple review cycles.

2022 - 2024

Education

Work History