Yaejin Cho

Key Skills

Software

Data Annotation Tech

Other

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Text

Top Task Types

Prompt Response Writing SFT

Question Answering

Text Generation

Translation Localization

Freelancer Overview

With hands-on experience in multilingual AI data annotation, I specialize in evaluating and refining AI-generated content in both English and Korean. My work focuses on comparative analysis, tone and intent assessment, and linguistic accuracy—ensuring high-quality data that improves large language model (LLM) outputs. I've contributed to the quality assurance pipeline by annotating diverse prompts, identifying cultural and syntactic inconsistencies, and applying contextual reasoning to enhance model performance. What sets me apart is my unique background in simultaneous interpretation and brand copywriting, which deepens my sensitivity to language nuance, interpersonal tone, and content adaptation across varied domains. This dual expertise allows me to provide highly context-aware feedback, align data with real-world use cases, and maintain consistency and clarity under tight deadlines. My strengths lie in multilingual fluency, stylistic judgment, and a sharp eye for detail—enabling me to deliver annotation work that is both precise and deeply informed by real communication dynamics.

Labeling Experience

Question Answering Accuracy Review

Data Annotation TechTextQuestion AnsweringEvaluation Rating

Evaluated AI responses to questions for factual accuracy and completeness. Annotated supporting context, flagged hallucinations, and rated outputs based on clarity and informativeness. Strengthened the QA capabilities of LLMs used in knowledge domains.

Evaluated AI responses to questions for factual accuracy and completeness. Annotated supporting context, flagged hallucinations, and rated outputs based on clarity and informativeness. Strengthened the QA capabilities of LLMs used in knowledge domains.

2024 - 2022

Prompt + Response Writing for Supervised Fine-Tuning (SFT)

Data Annotation TechTextPrompt Response Writing SFT

Authored high-quality prompts and completions tailored for real-world tasks and user needs. Prioritized natural tone, diversity, and task relevance. Supported SFT pipelines for improved model alignment and helpfulness.

Authored high-quality prompts and completions tailored for real-world tasks and user needs. Prioritized natural tone, diversity, and task relevance. Supported SFT pipelines for improved model alignment and helpfulness.

2024

AI Output Evaluation & Rating

Data Annotation TechTextEvaluation Rating

Performed comparative evaluations of LLM responses, rating them based on fluency, helpfulness, and factual integrity. Flagged biased or off-tone outputs. Played a key role in feedback loops for model improvement.

Performed comparative evaluations of LLM responses, rating them based on fluency, helpfulness, and factual integrity. Flagged biased or off-tone outputs. Played a key role in feedback loops for model improvement.

2024

Summarization Task Evaluation

Data Annotation TechTextEvaluation Rating

Assessed generated summaries for conciseness, relevance, and accuracy. Ensured coverage of key points and avoidance of hallucinated facts. Provided structured feedback to enhance model summarization abilities.

Assessed generated summaries for conciseness, relevance, and accuracy. Ensured coverage of key points and avoidance of hallucinated facts. Provided structured feedback to enhance model summarization abilities.

2024

Emotion Recognition in Conversations

Data Annotation TechTextRelationshipClassification

Tagged emotional signals in user conversations, identifying nuanced tones such as comfort, frustration, or hope. Labeled multiple emotion types per dialogue, contributing to emotionally intelligent LLM responses for wellness and support scenarios.

Tagged emotional signals in user conversations, identifying nuanced tones such as comfort, frustration, or hope. Labeled multiple emotion types per dialogue, contributing to emotionally intelligent LLM responses for wellness and support scenarios.

2024

Show more

Education

Work History