For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
Kyoko Cihlar Okuno

Kyoko Cihlar Okuno

JA/EN Bilingual LLM Evaluator & RLHF Data Annotation Specialist

Australia flagSydney, Australia
$30.00/hrEntry LevelOther

Key Skills

Software

Other

Top Subject Matter

No subject matter listed

Top Data Types

AudioAudio
DocumentDocument
TextText

Top Task Types

Audio Recording
Prompt Response Writing SFT
RLHF

Freelancer Overview

I am a bilingual Japanese-English AI Trainer and Data Annotation Specialist with hands-on experience in evaluating and enhancing large language models (LLMs). My expertise includes RLHF (Reinforcement Learning with Human Feedback), SFT (Supervised Fine-Tuning), and high-quality prompt and response evaluation for both text and audio data. I have worked on conversational AI outputs, metadata labeling, and ensuring AI systems deliver accurate, contextually appropriate responses. In addition, I have professional experience in financial operations and client service, giving me a solid understanding of finance-related content. This background allows me to contribute effectively to AI training projects in financial domains, as well as general-purpose NLP and audio tasks.

Entry LevelEnglishJapanese

Labeling Experience

LLM Conversational AI Evaluation & Response Assessment

OtherTextRLHFEvaluation Rating
Conducted Japanese-language conversational AI evaluation tasks with persona-based LLM chatbots. For each prompt, two model responses were generated, and I engaged in 12-turn conversations with the preferred response. Each turn was evaluated, selecting the better response and providing a short reason in English. Errors during conversation were submitted and reported. Tasks required detailed reasoning, accuracy assessment, and consistency across multiple conversation turns. All work was performed using the Outlier Platform.

Conducted Japanese-language conversational AI evaluation tasks with persona-based LLM chatbots. For each prompt, two model responses were generated, and I engaged in 12-turn conversations with the preferred response. Each turn was evaluated, selecting the better response and providing a short reason in English. Errors during conversation were submitted and reported. Tasks required detailed reasoning, accuracy assessment, and consistency across multiple conversation turns. All work was performed using the Outlier Platform.

2025 - 2025

Japanese Audio Prompt Creation & Annotation

OtherAudioClassificationEvaluation Rating
I created Japanese audio prompts based on a structured set of main and sub-categories, generating a wide range of prompt types according to detailed guidelines. Using the Outlier platform, I recorded audio samples and performed comprehensive annotation, including documenting noise levels, background sounds, tone, and all detectable non-speech audio elements. I also reviewed the model’s transcription output for accuracy and corrected any errors to ensure high-quality, guideline-aligned data. Tasks required attention to detail, consistency, and precise linguistic and acoustic evaluation.

I created Japanese audio prompts based on a structured set of main and sub-categories, generating a wide range of prompt types according to detailed guidelines. Using the Outlier platform, I recorded audio samples and performed comprehensive annotation, including documenting noise levels, background sounds, tone, and all detectable non-speech audio elements. I also reviewed the model’s transcription output for accuracy and corrected any errors to ensure high-quality, guideline-aligned data. Tasks required attention to detail, consistency, and precise linguistic and acoustic evaluation.

2025 - 2025

Japanese Conversational AI Evaluation & Meta-Review

OtherTextClassificationText Generation
I evaluated Japanese-language conversational interactions between AI models and users, focusing on fluency, tone, and locale appropriateness. In addition, I performed meta-review tasks by assessing English-language grader evaluations to verify scoring accuracy and ensure consistency with detailed guidelines. My responsibilities included reviewing multi-turn conversations, identifying errors, providing justification for evaluation decisions in English, and ensuring high-quality, guideline-aligned assessments across all assigned tasks.

I evaluated Japanese-language conversational interactions between AI models and users, focusing on fluency, tone, and locale appropriateness. In addition, I performed meta-review tasks by assessing English-language grader evaluations to verify scoring accuracy and ensure consistency with detailed guidelines. My responsibilities included reviewing multi-turn conversations, identifying errors, providing justification for evaluation decisions in English, and ensuring high-quality, guideline-aligned assessments across all assigned tasks.

2025 - 2025

Japanese Conversational AI Evaluation & Audio/Text Data Annotation

OtherAudioRLHFEvaluation Rating
Conducted Japanese conversational AI tasks using AI models installed on mobile devices. Generated conversations by inputting predetermined prompts, recorded both screen and audio, and saved conversation transcripts on the mobile device. Recorded videos were uploaded through a dedicated computer-based submission system. A separate evaluation platform was used to review each conversation in detail—assessing accuracy, naturalness, and adherence to instructions, and determining which model performed better. All evaluation and reasoning were completed in English.

Conducted Japanese conversational AI tasks using AI models installed on mobile devices. Generated conversations by inputting predetermined prompts, recorded both screen and audio, and saved conversation transcripts on the mobile device. Recorded videos were uploaded through a dedicated computer-based submission system. A separate evaluation platform was used to review each conversation in detail—assessing accuracy, naturalness, and adherence to instructions, and determining which model performed better. All evaluation and reasoning were completed in English.

2025 - 2025

Education

T

TAFE St George College

Diploma of Financial Planning, Financial Planning

Diploma of Financial Planning
2014 - 2014
S

Seijo University Junior College

Associate in Arts, Arts

Associate in Arts
1992 - 1992

Work History

C

CBD Flowers

Sales & Florist Assistant

Sydney
2022 - 2025
D

Deutsche Bank

Settlement Officer / Team Leader

Sydney
2006 - 2011