Kei Omote

Key Skills

Software

Clickworker

Data Annotation Tech

Labelbox

Lionbridge

Scale AI

Telus

Don't disclose

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Audio Recording

Prompt Response Writing SFT

Red Teaming

RLHF

Translation Localization

Freelancer Overview

I have hands-on experience in AI training and data labeling, specializing in Japanese and English linguistic tasks. My work has included LLM prompt engineering, output evaluation, and instruction writing for generative AI models. I’ve contributed to projects involving relevance scoring, translation quality assessment, and tone and register evaluation for both conversational and task-oriented prompts. My background in linguistics and bilingual fluency allows me to detect subtle errors in grammar, context, and cultural appropriateness, which is essential for refining model outputs. What sets me apart is my ability to balance linguistic precision with cultural nuance, ensuring that AI systems perform accurately across both Japanese and English use cases. I am detail-oriented, adaptable to evolving guidelines, and experienced in annotation tools and structured feedback systems. My cross-cultural insights and strong written communication skills make me especially effective in multilingual NLP and localization-focused tasks.

Labeling Experience

Cultural Awareness Prompt Writing

Don T DiscloseTextRed TeamingPrompt Response Writing SFT

Contributed to a global Cultural Awareness Prompt Writing project aimed at evaluating AI models for cultural sensitivity and linguistic accuracy. Created diverse, high-quality prompts targeting specific cultural attributes (e.g., norms, language use, naming conventions) across multiple complexity levels. Applied in-depth regional knowledge of Japanese and American culture to design realistic, context-rich queries that test AI understanding of tone, formality, social customs, and intra-cultural variation.

Contributed to a global Cultural Awareness Prompt Writing project aimed at evaluating AI models for cultural sensitivity and linguistic accuracy. Created diverse, high-quality prompts targeting specific cultural attributes (e.g., norms, language use, naming conventions) across multiple complexity levels. Applied in-depth regional knowledge of Japanese and American culture to design realistic, context-rich queries that test AI understanding of tone, formality, social customs, and intra-cultural variation.

2024

Project Utterance

LabelboxAudioData Collection

Contributed to the Utterance Labeling project by recording single-turn voice prompts on mobile devices in natural environments to train and evaluate AI speech models. Ensured recordings met project guidelines regarding accent authenticity, background noise, and clarity. Labeled metadata including speech difficulty, background activity type, and audio quality to support accurate model training for voice recognition and multilingual understanding tasks.

Contributed to the Utterance Labeling project by recording single-turn voice prompts on mobile devices in natural environments to train and evaluate AI speech models. Ensured recordings met project guidelines regarding accent authenticity, background noise, and clarity. Labeled metadata including speech difficulty, background activity type, and audio quality to support accurate model training for voice recognition and multilingual understanding tasks.

2024 - 2024

Cypher Evals

Scale AITextRLHFEvaluation Rating

On the Cypher_Evals project, I evaluated AI-generated responses to user prompts by rating them across multiple dimensions, including localization, instruction following, truthfulness, tone, and safety. I compared two outputs per prompt and provided detailed justifications and preference rankings based on nuanced qualitative analysis. Tasks required identifying prompt types, verifying reference text alignment, detecting factual inaccuracies, and applying cultural and linguistic context in Japanese and English.

On the Cypher_Evals project, I evaluated AI-generated responses to user prompts by rating them across multiple dimensions, including localization, instruction following, truthfulness, tone, and safety. I compared two outputs per prompt and provided detailed justifications and preference rankings based on nuanced qualitative analysis. Tasks required identifying prompt types, verifying reference text alignment, detecting factual inaccuracies, and applying cultural and linguistic context in Japanese and English.

2024 - 2024

Project Raven

Data Annotation TechTextRLHFEvaluation Rating

On the Raven Prompt Categorization project, I analyzed and labeled user-generated prompts by identifying their primary use cases (e.g., Brainstorming, Advice, Summarization, Creative Writing) based on context and intent. This required careful linguistic judgment to distinguish subtle differences between overlapping categories and ensure consistent classification. I applied a strong understanding of prompt design, user intent, and natural language to improve how AI models interpret and respond to diverse input types across English and Japanese tasks.

On the Raven Prompt Categorization project, I analyzed and labeled user-generated prompts by identifying their primary use cases (e.g., Brainstorming, Advice, Summarization, Creative Writing) based on context and intent. This required careful linguistic judgment to distinguish subtle differences between overlapping categories and ensure consistent classification. I applied a strong understanding of prompt design, user intent, and natural language to improve how AI models interpret and respond to diverse input types across English and Japanese tasks.

2023 - 2024

Education

Work History