Shingo Sato

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

RLHF

Freelancer Overview

I have over 18 months of intensive experience as an AI Training Specialist, specializing in the evaluation and refinement of Japanese Large Language Models (LLMs). My expertise lies in RLHF (Reinforcement Learning from Human Feedback), where I consistently perform high-level Response Ranking, Fact Verification, and Prompt Engineering. I excel at identifying and documenting complex failure modes, particularly in areas requiring deep linguistic precision, pragmatic nuances, and structural syntax. By grounding model responses in rigorous fact-checking and source-material analysis, I have contributed to hardening model reasoning and ensuring natural, professional-grade Japanese output. What sets me apart is the combination of an educational background from Osaka University of Foreign Studies and a perfect TOEIC score of 990, ensuring superior English-to-Japanese conceptual alignment. My work is characterized by over 20 years of professional experience in language education and content creation, which allows for a metacognitive approach to communication—focusing on the "why" behind model errors. This commitment to quality has been recognized, resulting in an invitation to take on a Managerial Role.

Labeling Experience

LLM Training and Linguistic Evaluation Specialist (Japanese)

Scale AITextRLHF

Specialized in high-stakes linguistic evaluation and RLHF for Japanese LLMs. - Response Ranking & Evaluation: Evaluated and ranked multiple AI-generated responses based on strict criteria, including helpfulness, truthfulness, and safety. - Linguistic Error Analysis: Identified and documented complex failure modes in Japanese text, such as subtle unnaturalness in honorifics, pragmatic inconsistencies, and syntactic errors. - Fact Verification & Grounding: Conducted rigorous fact-checking of model outputs against provided source materials to ensure factual accuracy and logical reasoning. - Creative Content Refinement: Performed "post-editing" on machine-generated content to transform robotic phrasing into natural, high-quality Japanese prose. Quality Leadership: Consistently delivered superior quality scores, leading to an invitation for a managerial role.

Specialized in high-stakes linguistic evaluation and RLHF for Japanese LLMs. - Response Ranking & Evaluation: Evaluated and ranked multiple AI-generated responses based on strict criteria, including helpfulness, truthfulness, and safety. - Linguistic Error Analysis: Identified and documented complex failure modes in Japanese text, such as subtle unnaturalness in honorifics, pragmatic inconsistencies, and syntactic errors. - Fact Verification & Grounding: Conducted rigorous fact-checking of model outputs against provided source materials to ensure factual accuracy and logical reasoning. - Creative Content Refinement: Performed "post-editing" on machine-generated content to transform robotic phrasing into natural, high-quality Japanese prose. Quality Leadership: Consistently delivered superior quality scores, leading to an invitation for a managerial role.

2024 - 2025

Education

Work History