Doyoon (Do)

Native Korean speaker · Daily LLM power user · Available for AI training & evaluation projects

Summary

Native Korean speaker with deep practical experience using large language models in daily creative and analytical workflows. Strong written Korean across registers (casual, formal, conversational), bilingual sensitivity (KO↔EN), and an editorial eye for nuance. Looking for AI training, evaluation, and Korean NLP projects.

Core Strengths

Relevant Experience

AI-Assisted Workflow Designer (Personal Practice)
2024 — Present
  • Built personalized AI agent setups using Claude Code, custom skills, and MCP servers for daily creative and research workflows.
  • Hands-on with prompt engineering, structured output design, and integrating LLMs into multi-step workflows.
  • Routine evaluation of model outputs in both Korean and English — directly familiar with where current LLMs succeed and fail in Korean (조사 오류, 부자연스러운 어순, 존댓말 일관성, 한국 문화 맥락 누락 등).
  • Maintain pectrono on GitHub for personal tools and experiments.

Skills

Languages Korean (Native) · English (Professional Working)
AI / LLM Claude · ChatGPT · Gemini · prompt engineering · model output evaluation · RLHF-style annotation · structured output design
Tools Claude Code · Obsidian · Git / GitHub · Notion · basic HTML / CSS / JS
Editorial Korean copywriting · KO↔EN translation review · tone & register editing · content quality evaluation

Interests

Korean NLP · LLM evaluation · creative AI applications.