Robert Kipngeno

Key Skills

Software

CloudFactory

Appen

Top Subject Matter

STEM (science, technology, engineering, mathematics)

Technical Reasoning

Technical Articles

Top Data Types

Text

Document

Top Task Types

RLHF

Freelancer Overview

Dedicated Data Associate with 2+ years of experience specializing in RLHF (Reinforcement Learning from Human Feedback) and high-complexity technical auditing. I possess a deep expertise in evaluating LLM outputs for logical consistency, factual accuracy, and adherence to intricate system constraints. My background in structured analysis and technical reasoning allows me to effectively audit multi-step "Chain-of-Thought" (CoT) prompts, ensuring that model responses are not only helpful and safe but mathematically and logically sound. Data Labeling/AI Training Experience My project portfolio includes lead roles in STEM-based data verification and technical fact-checking for Tier-1 model developers. I have successfully executed large-scale tasks on platforms like CloudFactory and Appen, specifically focusing on ranking multi-turn conversations and performing deep-dive audits of AI-generated technical summaries. By identifying subtle hallucinations and verifying claims against primary documentation, I consistently maintain a 98.5% quality accuracy rating while delivering high-density training data for complex reasoning models.

Labeling Experience

Lead Data Associate (Freelance)

CloudfactoryTextRLHF

As a Lead Data Associate at CloudFactory, I performed high-volume RLHF tasks involving technical reasoning and STEM-based prompts. My core responsibility was auditing and correcting AI-generated text to align with detailed system instructions and factual data. I played a key role in minimizing technical inaccuracies and improving output quality for a major LLM developer. • Verified and evaluated 500+ AI responses daily in multi-turn conversations. • Focused on STEM, technical, and logical reasoning subject matter. • Collaborated on projects aimed at reducing AI 'hallucinations' in documentation. • Maintained a 98.5% quality score throughout the evaluation process.

As a Lead Data Associate at CloudFactory, I performed high-volume RLHF tasks involving technical reasoning and STEM-based prompts. My core responsibility was auditing and correcting AI-generated text to align with detailed system instructions and factual data. I played a key role in minimizing technical inaccuracies and improving output quality for a major LLM developer. • Verified and evaluated 500+ AI responses daily in multi-turn conversations. • Focused on STEM, technical, and logical reasoning subject matter. • Collaborated on projects aimed at reducing AI 'hallucinations' in documentation. • Maintained a 98.5% quality score throughout the evaluation process.

2025 - Present

AI Content Evaluator

AppenText

As an AI Content Evaluator at Appen, I was responsible for assessing AI-generated summaries and search results for a global search platform. Tasks included ranking relevance, identifying factual errors, and detecting biased or sensitive content in the data. This work helped enhance both the quality and safety of machine learning training data. • Evaluated long-form technical article summaries for accuracy and brevity. • Performed sensitivity and bias checks on training data outputs. • Wrote detailed justifications for ranking decisions and flagged logical fallacies. • Contributed to safety improvements in the model training process.

As an AI Content Evaluator at Appen, I was responsible for assessing AI-generated summaries and search results for a global search platform. Tasks included ranking relevance, identifying factual errors, and detecting biased or sensitive content in the data. This work helped enhance both the quality and safety of machine learning training data. • Evaluated long-form technical article summaries for accuracy and brevity. • Performed sensitivity and bias checks on training data outputs. • Wrote detailed justifications for ranking decisions and flagged logical fallacies. • Contributed to safety improvements in the model training process.

2024 - 2024

Education

Work History