Open to opportunities

Yoshitaka Kurumada

@yoshitakakurumada

Message

AI/ML data evaluator focused on accurate search relevance and robust guideline QA.

Japan

Message

What I'm looking for

I’m looking for full-time work where I can evaluate AI outputs and training data, enforce clear guidelines, and improve search relevance through intent analysis and high-accuracy QA—ideally with growth in ML evaluation and LLM benchmarking.

I’m an AI/ML data evaluator supporting search quality through detailed qualitative justification, strict guideline adherence, and careful handling of ambiguous user intents. At Appen, I’ve consistently ensured the logic behind each evaluation is clear, robust, and aligned with project requirements.

I’ve evaluated video and photo relevance against user search queries to improve search engine algorithms. I analyze user search intent, categorize content by usefulness and engagement quality, and maintain performance above targets—exceeding 2500 videos per week while keeping weekly accuracy above 90%.

I also contribute to quality improvement by discovering conflicts in rating among members and documenting disputed elements, separating subjective reasoning from guideline-following logic. My work has led to selection as a candidate for Auditor (Quality Assurance) roles, reflecting consistent high performance and exceptional guideline adherence.

Experience

Work history, roles, and key accomplishments

Current

Music Search Rating Evaluator

Current

Appen

Apr 2025 - Present (1 year 2 months)

Provided detailed qualitative justification for music search ratings, resolving ambiguous user intent per project guidelines and ensuring robust, guideline-aligned decisions. Selected for Auditor (Quality Assurance) roles based on consistently high performance and exceptional guideline adherence.

Data Labeling Content Moderation ChatGpt Claude

Search Video/Photo Evaluator

Appen

Jan 2026 - Feb 2026 (1 month)

Evaluated video content relevance to user search queries by analyzing search intent and categorizing results by usefulness, accuracy, and engagement quality. Applied complex and evolving guidelines to produce high-quality training data, maintaining 90%+ weekly accuracy and contributing to conflict resolution between members.

Video Photo Relevance Assessment User Intent Data Labeling Guideline QA Accuracy Tracking

Search Video/Photo Evaluator

Appen

May 2025 - Dec 2025 (7 months)

Evaluated video content relevance to user search queries by analyzing search intent and categorizing results by usefulness, accuracy, and engagement quality. Exceeded performance targets by 2,500+ videos/week while maintaining 90%+ weekly accuracy, dedicating 40+ hours/week to moderation and data evaluation.

Content Relevance Evaluation Result Usefulness Assessment Data Labeling Guideline Based QA Content Moderation Accuracy Tracking