Semin Kim

Key Skills

Software

Clickworker

Data Annotation Tech

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Audio Recording

Classification

Evaluation Rating

Fine Tuning

Translation Localization

Freelancer Overview

I have over a year of experience in data labeling and AI training data projects through platforms such as Outlier (Scale AI) and DataAnnotation. At Outlier, I primarily worked on math-related tasks, leveraging my academic background to generate and evaluate prompts and responses. My proficiency in LaTeX allowed me to clearly and accurately format complex mathematical expressions, contributing to higher quiality outputs. Due to the high quality and consistency of my work, I was promoted to a reviewer role in the majority of the projects I participated in. On the DataAnnotation Platform, I focused on evaluating model outputs in Korean, including instruction-following and comparative assessments. I consistently high-quality work and worked as a reviewer across multiple projects. I continue to contribute actively to both platforms, maintaining a strong track record of accuracy and adherence to task quidelines.

Labeling Experience

L10 Reviewer

Scale AIAudioClassificationAudio Recording

Generated and classified voice recordings across a variety of topics and environments. Provided voice recordings tailored to different topics and locations, and was promoted to L0 reviewer based on consistent quality. Later advanced to L10 reviewer. As a reviewer, I evaluated the transcription accuracy of others' recordings and classified the recording environment, topic category, and noise level.

Generated and classified voice recordings across a variety of topics and environments. Provided voice recordings tailored to different topics and locations, and was promoted to L0 reviewer based on consistent quality. Later advanced to L10 reviewer. As a reviewer, I evaluated the transcription accuracy of others' recordings and classified the recording environment, topic category, and noise level.

2025

Data Annotator

Data Annotation TechTextClassificationEvaluation Rating

Developed 5-10 quaility criteria, including implicit, explicit, subjective, and objective measures, to evaluate and improve model reponses. Provided clear and detailed evalutation criteria to optimize model responses to user prompts. Assessed model outputs based on these criteria and suggested improvements to enhance clarity, accuracy, and relevance.

Developed 5-10 quaility criteria, including implicit, explicit, subjective, and objective measures, to evaluate and improve model reponses. Provided clear and detailed evalutation criteria to optimize model responses to user prompts. Assessed model outputs based on these criteria and suggested improvements to enhance clarity, accuracy, and relevance.

2025 - 2025

AI Data Annotator

Data Annotation TechImageClassificationEvaluation Rating

Created image-based prompts across various categories and evaluated the model's responses. Sourced or generated images by category and crafted promts according to task specific guidelines. Evaluated model outputs based on multiple criteria, and later reviewed and provided feedback on work submitted by other contributors.

Created image-based prompts across various categories and evaluated the model's responses. Sourced or generated images by category and crafted promts according to task specific guidelines. Evaluated model outputs based on multiple criteria, and later reviewed and provided feedback on work submitted by other contributors.

2025 - 2025

Reviewer

Scale AITextClassificationEvaluation Rating

Scope of the project: Collected mathematical prompts and answers designed to challenge and outperform the model Specific Data Labeling Tasks Performed: Generated mathematical prompts and responses ranging from grad school to graduate level. Promoted to reviewer role, where responsibilities included classifying prompts submitted by other attempters and evaluating the model's responses against attempter's answers. Project size: global (participated in the Korean language section) Quality measures: Maintained over 4 out of 5 quality level

Scope of the project: Collected mathematical prompts and answers designed to challenge and outperform the model Specific Data Labeling Tasks Performed: Generated mathematical prompts and responses ranging from grad school to graduate level. Promoted to reviewer role, where responsibilities included classifying prompts submitted by other attempters and evaluating the model's responses against attempter's answers. Project size: global (participated in the Korean language section) Quality measures: Maintained over 4 out of 5 quality level

2025 - 2025

Education

Work History