Confidence Osaze Aigbedion

Key Skills

Software

Scale AI

Top Subject Matter

LLM Evaluation and Safety

Multimodal Data Annotation (Text, Image, Computer Vision)

LLM Reasoning and Safety Evaluation

Top Data Types

Text

Image

Audio

Document

Top Task Types

Prompt + Response Writing (SFT)

Segmentation

RLHF

Red Teaming

Computer Programming/Coding

Freelancer Overview

AI Safety & Prompt Engineering Specialist. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Scale AI. Education includes Master of Science, Universität Trier (2025) and Professional Certification, ALX (2024). AI-training focus includes data types such as Text, Image, and Audio and labeling workflows including Prompt + Response Writing (SFT), Segmentation, and RLHF.

Labeling Experience

Multimodal Data Evaluator

ImageSegmentation

As a Multimodal Data Evaluator, I led annotation projects involving the labeling of text, images, and computer-vision trajectories to develop high-quality training data for AI agents. I recorded and annotated workflows in Linux environments to support human-in-the-loop tasks for training AI to navigate software interfaces. Collaborating directly with project managers, I refined labeling guidelines for client-specific quality benchmarks. • Executed multimodal annotation tasks including text, image, and computer vision labeling. • Used Linux desktop setups to support human-in-the-loop recording and annotation. • Refined and improved data quality metrics and labeling guidelines through collaboration. • Raised data annotation standards for specialized client AI systems.

As a Multimodal Data Evaluator, I led annotation projects involving the labeling of text, images, and computer-vision trajectories to develop high-quality training data for AI agents. I recorded and annotated workflows in Linux environments to support human-in-the-loop tasks for training AI to navigate software interfaces. Collaborating directly with project managers, I refined labeling guidelines for client-specific quality benchmarks. • Executed multimodal annotation tasks including text, image, and computer vision labeling. • Used Linux desktop setups to support human-in-the-loop recording and annotation. • Refined and improved data quality metrics and labeling guidelines through collaboration. • Raised data annotation standards for specialized client AI systems.

2025 - Present

AI Safety & Prompt Engineering Specialist

TextPrompt Response Writing SFT

As an AI Safety & Prompt Engineering Specialist, I designed and evaluated prompt-testing frameworks to assess the reasoning and ethical alignment of large language models. I crafted creative adversarial and dialogue-driven prompts, performed linguistic and sentiment analysis, and authored evaluation reports highlighting issues and recommendations. My work focused on strengthening model safety and identifying high-risk behavioral triggers to mitigate vulnerabilities. • Developed prompt chains to assess LLM resilience against jailbreaks and misinformation. • Conducted linguistic and sentiment analysis to measure model reliability and ethical adherence. • Delivered detailed evaluation reports specifying bias triggers and mitigation approaches. • Enhanced LLM safety by recommending specific improvements based on model weaknesses.

As an AI Safety & Prompt Engineering Specialist, I designed and evaluated prompt-testing frameworks to assess the reasoning and ethical alignment of large language models. I crafted creative adversarial and dialogue-driven prompts, performed linguistic and sentiment analysis, and authored evaluation reports highlighting issues and recommendations. My work focused on strengthening model safety and identifying high-risk behavioral triggers to mitigate vulnerabilities. • Developed prompt chains to assess LLM resilience against jailbreaks and misinformation. • Conducted linguistic and sentiment analysis to measure model reliability and ethical adherence. • Delivered detailed evaluation reports specifying bias triggers and mitigation approaches. • Enhanced LLM safety by recommending specific improvements based on model weaknesses.

2025 - Present

AI Content Evaluator (RLHF Specialist)

Scale AITextRLHF

As an AI Content Evaluator (RLHF Specialist), I evaluated LLM-generated content for logical consistency, reasoning accuracy, and safety using RLHF protocols. I annotated complex reasoning pathways and ranked model responses to support the development of gold-standard data for fine-tuning. My role demanded attention to reducing hallucinations and optimizing technical content responses for advanced AI systems. • Provided expert review and ratings for LLM outputs, focusing on logical accuracy and safety. • Annotated reasoning chains and created plausible alternative trajectories for model training. • Collaborated on gold-standard data sets to enhance LLM's technical accuracy. • Utilized Linux environments to support text and limited multimodal annotations.

As an AI Content Evaluator (RLHF Specialist), I evaluated LLM-generated content for logical consistency, reasoning accuracy, and safety using RLHF protocols. I annotated complex reasoning pathways and ranked model responses to support the development of gold-standard data for fine-tuning. My role demanded attention to reducing hallucinations and optimizing technical content responses for advanced AI systems. • Provided expert review and ratings for LLM outputs, focusing on logical accuracy and safety. • Annotated reasoning chains and created plausible alternative trajectories for model training. • Collaborated on gold-standard data sets to enhance LLM's technical accuracy. • Utilized Linux environments to support text and limited multimodal annotations.

2024 - 2024

Quality Assurance Specialist

Audio

As a Quality Assurance Specialist, I directed the QA lifecycle for digital audio and text datasets used in NLP model training. I provided actionable feedback to data collection teams, improving the quality, consistency, and usability of training data. Comprehensive documentation on edge-case detection and quality control ensured continuous improvement within evaluation frameworks. • Oversaw dataset evaluation processes for compliance with NLP model requirements. • Enhanced training data integrity through detailed feedback and consistency controls. • Drafted documentation to improve edge-case awareness and resolution. • Ensured high-quality audio and text datasets for downstream AI/NLP applications.

As a Quality Assurance Specialist, I directed the QA lifecycle for digital audio and text datasets used in NLP model training. I provided actionable feedback to data collection teams, improving the quality, consistency, and usability of training data. Comprehensive documentation on edge-case detection and quality control ensured continuous improvement within evaluation frameworks. • Oversaw dataset evaluation processes for compliance with NLP model requirements. • Enhanced training data integrity through detailed feedback and consistency controls. • Drafted documentation to improve edge-case awareness and resolution. • Ensured high-quality audio and text datasets for downstream AI/NLP applications.

2022 - 2023

Education

Work History