Elmars

Key Skills

Software

CVAT

Data Annotation Tech

Label Studio

Mercor

Internal/Proprietary Tooling

Top Subject Matter

Content evaluation

information verification

and AI response/content quality assessment

Top Data Types

Text

Audio

Image

Top Task Types

RLHF

Question Answering

Evaluation/Rating

Freelancer Overview

I am a detail-oriented Computer Science graduate with experience in content evaluation, research, and structured feedback writing, skills that closely align with AI training data and data labeling work. Through independent remote projects, I reviewed written materials and online content to identify issues with accuracy, consistency, and quality, and provided organized, actionable feedback. I also conducted internet-based research, clearly summarized findings, and maintained a high level of attention to detail with minimal supervision. My strong critical reading and analytical thinking abilities enable me to evaluate information carefully and ensure high-quality outputs. In addition to my academic background in Computer Science from Michael Okpara University of Agriculture, Umudike, I have developed strong familiarity with digital tools, online platforms, and technical documentation through coursework and collaborative technology projects. I am passionate about artificial intelligence, language models, and AI response evaluation, with a particular interest in improving the quality of AI-generated content. My combination of structured communication, problem-solving ability, research skills, and independent remote work experience sets me apart as a reliable contributor for AI training data, annotation, and evaluation tasks.

Labeling Experience

Audio quality

Don't discloseAudioRLHF

I previously worked on AI data evaluation and ranking projects involving audio and response-quality assessment tasks similar to reinforcement learning from human feedback (RLHF). One example was an audio-quality pairwise evaluation project where I compared AI-generated audio outputs, assessed clarity, accuracy, naturalness, pronunciation, and overall user experience, then selected the better-performing response based on predefined quality guidelines. The work also involved identifying inconsistencies, transcription issues, low-quality generations, and edge cases while providing structured feedback to improve model performance.

2025 - 2026

Education

Work History