Clement Ogol

Key Skills

Software

Internal/Proprietary Tooling

CloudFactory

Top Subject Matter

No subject matter listed

Top Data Types

Video

Audio

Image

Computer Code Programming

Top Task Types

RLHF

Evaluation Rating

Classification

Emotion Recognition

Bounding Box

Fine Tuning

Freelancer Overview

Senior AI Trainer and Data Analyst specializing in RLHF (Reinforcement Learning from Human Feedback) and Generative AI evaluation. I have extensive experience in high-complexity data labeling for Video and Audio modalities, specifically in distinguishing AI hallucinations and physics glitches. Proven track record of maintaining 98%+ quality accuracy and auditing large-scale datasets for Computer Vision and NLP projects. Proficient in Python and SQL for technical data validation.

Labeling Experience

AI Trainer - Agentic Workflow Automation

Internal Proprietary ToolingComputer Code ProgrammingRLHFFine Tuning

Generated training datasets for Agentic AI models designed to execute complex desktop tasks autonomously. Created step-by-step action traces for human-computer interaction workflows, such as web development (coding and deploying websites) and administrative automation (email scheduling and CRM management). Demonstrated logic paths to teach the AI how to navigate UIs and execute multi-step commands like a human user.

Generated training datasets for Agentic AI models designed to execute complex desktop tasks autonomously. Created step-by-step action traces for human-computer interaction workflows, such as web development (coding and deploying websites) and administrative automation (email scheduling and CRM management). Demonstrated logic paths to teach the AI how to navigate UIs and execute multi-step commands like a human user.

2025 - 2025

Audio Sentiment & Emotion Recognition

Internal Proprietary ToolingAudioClassificationEmotion Recognition

Analyzed audio clips to create a Ground Truth dataset for Voice Emotion Recognition algorithms. Labeled subtle vocal cues and emotional intonations (sarcasm, urgency, joy) to optimize NLP models. Refined categorization rules for ambiguous audio data.

Analyzed audio clips to create a Ground Truth dataset for Voice Emotion Recognition algorithms. Labeled subtle vocal cues and emotional intonations (sarcasm, urgency, joy) to optimize NLP models. Refined categorization rules for ambiguous audio data.

2023 - 2025

RLHF Video Generative AI Evaluation

Internal Proprietary ToolingVideoRLHFEvaluation Rating

Evaluated the output of generative video models, applying strict ground truth standards to distinguish between high-fidelity footage and AI hallucinations. Identified complex video artifacts such as physics glitches, morphing errors, and temporal inconsistencies. Provided structured technical feedback to engineering teams for model retraining and audited junior annotator work to ensure lighting and texture consistency.

Evaluated the output of generative video models, applying strict ground truth standards to distinguish between high-fidelity footage and AI hallucinations. Identified complex video artifacts such as physics glitches, morphing errors, and temporal inconsistencies. Provided structured technical feedback to engineering teams for model retraining and audited junior annotator work to ensure lighting and texture consistency.

2023 - 2025

Computer Vision Data Annotator (2D & 3D)

CloudfactoryImageBounding BoxClassification

Performed high-precision 2D and 3D data annotation for computer vision models. Specialized in labeling complex datasets using bounding boxes and 3D cuboids (LiDAR) for autonomous driving and robotics applications. Ensured strict adherence to spatial accuracy guidelines, handling complex scenarios like object occlusion, truncation, and multi-sensor fusion.

Performed high-precision 2D and 3D data annotation for computer vision models. Specialized in labeling complex datasets using bounding boxes and 3D cuboids (LiDAR) for autonomous driving and robotics applications. Ensured strict adherence to spatial accuracy guidelines, handling complex scenarios like object occlusion, truncation, and multi-sensor fusion.

2021 - 2022

Education

Work History