Sangwoo Yu

Key Skills

Software

Labelbox

Scale AI

Other

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Text

Top Task Types

Audio Recording

Computer Programming Coding

Data Collection

Question Answering

Text Generation

Freelancer Overview

Experienced AI Training Contributor with a strong background in prompt engineering, data annotation, and LLM evaluation. I have participated in multiple projects involving Korean voice recording, natural dialogue generation, and dataset review for large language model development. My work includes designing task-specific prompts, generating high-quality text responses, and providing detailed feedback on model outputs. I have also compared and evaluated models such as ChatGPT and Gemini Pro 2.5, offering structured performance reviews. In addition, I contributed UI evaluation feedback and code suggestions to improve model interaction environments. With a keen eye for language quality and data integrity, I bring precision and consistency to every AI training task.

Labeling Experience

Korean LLM Data Generation & Evaluation

Scale AITextText SummarizationEvaluation Rating

Collaborated on multilingual AI training by generating high-quality Korean language data through scripted and unscripted conversations. Designed prompts and generated responses tailored to diverse AI use cases. Evaluated the quality of model outputs for semantic accuracy, clarity, and tone. Participated in front-end UI review tasks to improve model usability and user experience. Consistently followed detailed project-specific guidelines and contributed to fine-tuning LLMs with culturally and linguistically accurate datasets.

Collaborated on multilingual AI training by generating high-quality Korean language data through scripted and unscripted conversations. Designed prompts and generated responses tailored to diverse AI use cases. Evaluated the quality of model outputs for semantic accuracy, clarity, and tone. Participated in front-end UI review tasks to improve model usability and user experience. Consistently followed detailed project-specific guidelines and contributed to fine-tuning LLMs with culturally and linguistically accurate datasets.

2024

LLM Response Evaluation & Text Annotation in Korean

LabelboxTextText GenerationEvaluation Rating

Worked on Korean-language evaluation tasks for large language models, comparing outputs from ChatGPT and Gemini Pro 2.5. Labeled AI-generated responses based on intent, naturalness, and accuracy, and identified errors or ambiguity to improve dataset quality. Created high-quality responses based on provided instructions and prompts, contributing to supervised fine-tuning (SFT) and alignment efforts. Maintained strict adherence to annotation guidelines and performance benchmarks.

Worked on Korean-language evaluation tasks for large language models, comparing outputs from ChatGPT and Gemini Pro 2.5. Labeled AI-generated responses based on intent, naturalness, and accuracy, and identified errors or ambiguity to improve dataset quality. Created high-quality responses based on provided instructions and prompts, contributing to supervised fine-tuning (SFT) and alignment efforts. Maintained strict adherence to annotation guidelines and performance benchmarks.

2025 - 2025

Education

Work History