Joy Kim

Key Skills

Software

OneForma

Other

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Audio Recording

Evaluation Rating

Text Generation

Translation Localization

Freelancer Overview

I am a dedicated data specialist with extensive experience in transcription, annotation, and quality assurance for AI training projects. Over the past several years, I have collaborated with global AI companies on Korean-English transcription, code-mix evaluation, and sentiment analysis tasks. My strong attention to detail, ability to follow complex guidelines, and bilingual fluency in Korean and English allow me to deliver consistently accurate results in fast-paced environments. In addition to language-focused work, I bring a background in accounting and business processes, which strengthens my ability to evaluate financial or technical content with precision. I have supported projects involving LLM evaluation, content quality review, and AI safety tasks, where I was responsible for scoring model outputs, flagging policy violations, and ensuring cultural relevance. With proven reliability, a flexible schedule, and strong technical skills, I am well positioned to contribute to high-quality AI training data across diverse industries.

Labeling Experience

Prompt creation and response evaluation

OtherTextQuestion AnsweringEvaluation Rating

Client: Outlier Project: Prompt Creation and Response Evaluation Duration: November 2024 – Present Platform: Data Compute / In-house Platform Factuality Evaluation Project Assessed the factual accuracy of two responses generated from a given prompt by cross-checking information with official and reliable sources, and provided supporting reference links as evidence. Period: December 2024 – June 2025 Rubric Creation Project Developed detailed evaluation rubrics based on topics in economics, society, education, and healthcare, and applied those rubrics to assess generated responses. Period: January 2025 – May 2025 Audio Recording Project Recorded speech based on selected topics across various fields, using either a knowledge-based or conversational style. Recordings were conducted both indoors and outdoors depending on noise levels, and included both individual and team-based sessions. Duration: December 2025 – Present

Client: Outlier Project: Prompt Creation and Response Evaluation Duration: November 2024 – Present Platform: Data Compute / In-house Platform Factuality Evaluation Project Assessed the factual accuracy of two responses generated from a given prompt by cross-checking information with official and reliable sources, and provided supporting reference links as evidence. Period: December 2024 – June 2025 Rubric Creation Project Developed detailed evaluation rubrics based on topics in economics, society, education, and healthcare, and applied those rubrics to assess generated responses. Period: January 2025 – May 2025 Audio Recording Project Recorded speech based on selected topics across various fields, using either a knowledge-based or conversational style. Recordings were conducted both indoors and outdoors depending on noise levels, and included both individual and team-based sessions. Duration: December 2025 – Present

2024 - 2025

Korean audio transcription project

OtherAudioSegmentationText Generation

I have worked on various audio transcription projects at Sigma AI, using the Data Compute platform. Audio Transcription Project Listened to Korean audio, segmented it by speaker, selected labels to distinguish speech from non-speech, and transcribed the spoken content accurately into Korean text. Duration: February 2024 – Present Emotion Classification Project Identified and categorized emotions and intentions perceived in the audio. Period: February 2025 – Present Preference Project Listened to two audio samples, selected the more natural one, and explained the reason. Period: May 2025 – Present Audio Correction Project Edited the given text to match the audio accurately. Period: July 2025 – Present

I have worked on various audio transcription projects at Sigma AI, using the Data Compute platform. Audio Transcription Project Listened to Korean audio, segmented it by speaker, selected labels to distinguish speech from non-speech, and transcribed the spoken content accurately into Korean text. Duration: February 2024 – Present Emotion Classification Project Identified and categorized emotions and intentions perceived in the audio. Period: February 2025 – Present Preference Project Listened to two audio samples, selected the more natural one, and explained the reason. Period: May 2025 – Present Audio Correction Project Edited the given text to match the audio accurately. Period: July 2025 – Present

2024

Native Korean Audio Transcriber – Hangul

LabelboxAudioText Generation

Client: OpenTrain Project: Korean Audio Transcription Project Duration: October 2025 – November 2025 Platform: Labelbox Responsibilities: Listened to short audio clips of approximately 30 seconds, identified and tagged vocal and non-vocal elements according to the guidelines, and transcribed Korean parts in Korean and English parts in English. After completing the transcription tasks, I also participated in reviewing the results. Over the two-month period, I worked approximately 120 hours.

Client: OpenTrain Project: Korean Audio Transcription Project Duration: October 2025 – November 2025 Platform: Labelbox Responsibilities: Listened to short audio clips of approximately 30 seconds, identified and tagged vocal and non-vocal elements according to the guidelines, and transcribed Korean parts in Korean and English parts in English. After completing the transcription tasks, I also participated in reviewing the results. Over the two-month period, I worked approximately 120 hours.

2025 - 2025

Korean-English CodeMix Transcription

OtherAudioTranslation Localization

Client: Sigma AI Project: Korean–English Bilingual Transcription and Reviewer Duration: April 2025 – October 2025 Platform: Data Compute Responsibilities: Listened to audio and transcribed clips while strictly applying a 0.1-second buffer rule to non-speech segments. Performed labeling of background noise, music, singing, and personally identifiable information (PII), as well as speaker intent classification. Maintained high accuracy by adhering to detailed project guidelines and quality metrics, while consistently meeting weekly targets (60–120 minutes of audio transcription per week). Also participated in QA reviews, providing corrections and feedback on other annotators’ work. This project supported the training of advanced multilingual speech recognition models and large language models (LLMs).

Client: Sigma AI Project: Korean–English Bilingual Transcription and Reviewer Duration: April 2025 – October 2025 Platform: Data Compute Responsibilities: Listened to audio and transcribed clips while strictly applying a 0.1-second buffer rule to non-speech segments. Performed labeling of background noise, music, singing, and personally identifiable information (PII), as well as speaker intent classification. Maintained high accuracy by adhering to detailed project guidelines and quality metrics, while consistently meeting weekly targets (60–120 minutes of audio transcription per week). Also participated in QA reviews, providing corrections and feedback on other annotators’ work. This project supported the training of advanced multilingual speech recognition models and large language models (LLMs).

2025 - 2025

Education

Work History