Ziyanda E.

Key Skills

Software

Internal/Proprietary Tooling

Top Subject Matter

LLM evaluation and AI output quality assurance

Technology & Digital Platforms

Software Quality Assurance (QA)

Top Data Types

Text

Image

Video

Top Task Types

RLHF

Red Teaming

Evaluation/Rating

Data Collection

Prompt + Response Writing (SFT)

Fine-tuning

Text Generation

Object Detection

Classification

Freelancer Overview

I worked as an Advanced AI Trainer, supporting large-scale AI training and data annotation projects focused on improving model accuracy, reasoning, and response quality. A big part of my role involved reviewing and labeling AI-generated outputs, comparing multiple responses, ranking them based on relevance and correctness, and identifying hallucinations, bias, formatting issues, and factual inconsistencies. I also worked extensively with prompt-response evaluation tasks, text classification, content moderation, and quality assurance reviews to ensure annotation standards were consistently met across datasets. Because many of the projects evolved quickly, I became very good at interpreting detailed guidelines, adapting to changing requirements, and making consistent judgment calls on nuanced language and edge-case scenarios. What really strengthened my experience was the quality-focused nature of the work. I wasn’t just labeling data mechanically; I was actively evaluating how models reasoned, where responses broke down, and how training data could be improved to produce more reliable outputs. I regularly handled high-volume review queues while maintaining strong accuracy metrics and turnaround times, and I often caught inconsistencies or guideline gaps that needed clarification before data could be finalized. My background in QA engineering also helped a lot because I naturally approach AI evaluation with a testing mindset: I pay attention to patterns, edge cases, logical failures, and overall output quality. That combination of AI training experience, analytical thinking, and detail-oriented review work has made me very comfortable working on complex annotation and model evaluation tasks.

Education

Work History