Francis Wanjohi

Key Skills

Software

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Text Generation

RLHF

Computer Programming Coding

Prompt Response Writing SFT

Freelancer Overview

I bring over 8 years of hands-on experience in data labeling, annotation, and AI training data, with a strong focus on quality assurance and technical evaluation. My expertise spans RLHF, LLM evaluation, and prompt engineering, where I have evaluated and ranked AI-generated code for correctness, efficiency, and security across domains such as operational systems, ERP (SAP, Oracle), and digital manufacturing workflows. I have conducted large-scale dataset labeling, authored "gold standard" responses for technical prompts, and performed rigorous audits to ensure data integrity and compliance in regulated environments. Skilled in Python, SQL, and advanced data validation, I am committed to delivering high-quality training data and process improvements that drive better AI model performance and reliable outcomes.

Labeling Experience

AI Technical Content & Code Dataset Labeling

Data Annotation TechTextText GenerationRLHF

This project involves the technical evaluation and optimization of frontier Large Language Models (LLMs) to improve their reasoning and coding capabilities. My role focuses on the "Gold Standard" training phase, where I evaluate and rank AI-generated code in Python, SQL, and JavaScript for functional correctness and logical efficiency. Key tasks include performing Reinforcement Learning from Human Feedback (RLHF) to identify and mitigate model hallucinations and logic bottlenecks. I also author high-quality, ground-truth code solutions and refine prompt engineering strategies to ensure models can handle complex, multi-step constraints and edge cases. Quality is maintained through strict adherence to rigorous style guidelines and a multi-step auditing process to ensure 100% data integrity for model fine-tuning.

This project involves the technical evaluation and optimization of frontier Large Language Models (LLMs) to improve their reasoning and coding capabilities. My role focuses on the "Gold Standard" training phase, where I evaluate and rank AI-generated code in Python, SQL, and JavaScript for functional correctness and logical efficiency. Key tasks include performing Reinforcement Learning from Human Feedback (RLHF) to identify and mitigate model hallucinations and logic bottlenecks. I also author high-quality, ground-truth code solutions and refine prompt engineering strategies to ensure models can handle complex, multi-step constraints and edge cases. Quality is maintained through strict adherence to rigorous style guidelines and a multi-step auditing process to ensure 100% data integrity for model fine-tuning.

2022 - 2023

Education

Work History