CV

Contact Information

Name Gahyun Yoo
Email padme0421@gmail.com

Experience

  • 2026 - 2026

    Agentic AI Engineer (Freelance)
    DeepAuto
    • Major conglomerate EPC automation project
    • Proposed & developed data lakehouse system for the EPC domain comprised of a tri-store knowledge base
    • Developed lakehouse-connected agents that each boosted recall for instrumentation drawing retrieval by 10%pp
    • Containerization for efficient and reliable forward deployment
    • Improved object detection by referencing project-specific legend sheets
  • 2025 - 2026

    Research Assistant
    SKI-ML Lab (PI: Jay-Yoon Lee), Seoul National University
    • [Research Project] Reward Model Finetuning for Diffusion Model Alignment Using Tree Structured Trajectories
    • [Government Project] Development of a Korean-English Translation Model for Safety Management of Pharmaceutical Products (Ministry of Food and Drug Safety): As project lead, developed a machine translation pipeline supporting an ensemble of inference strategies using in-context examples and dictionary entries, and instruction tuning & preference optimization training, outperforming several SOTA commercial models
    • [TA] Special Topic: Multimodal & Diffusion Models
    • [Coursework] Special Topic: Generative Modeling with Diffusion and Flow-Based Models
  • 2023 - 2025

    Research Intern
    SKI-ML Lab (PI: Jay-Yoon Lee), Seoul National University
    • [Research Project] Application of energy based training with quality estimators as score models for low resource machine translation, resulted in a publication at COLING 2025
    • [Research Project] Advised machine translation project, resulting in publication at ACL 2025
  • 2022 - 2022

    Research Intern
    NCSOFT NLP Center
    • Developed a decision tree classification algorithm for distinguishing literal and non-literal translation, using linguistic features, alignment score from text embedding similarity, and multilingual knowledge graphs (>90% accuracy on open source English-Chinese dataset)

Education

  • 2026 - 2028

    MS
    University of Southern California
    Computer Science
  • 2020 - 2025

    BS
    Seoul National University
    Linguistics & Computer Science

Publications

  • 2025
    Improving NMT Models by Retrofitting Quality Estimators into Trainable Energy Loss
    International Conference on Computational Linguistics
    Gahyun Yoo, Jay-Yoon Lee
  • 2025
    BridG MT: Enhancing LLMs' Machine Translation Capabilities with Sentence Bridging and Gradual MT
    Findings of the Association for Computational Linguistics
    Seung-Woo Choi, Gahyun Yoo, Jay-Yoon Lee
  • 2026
    Generalizing Numerical Reasoning in Table Data through Operation Sketches and Self-Supervised Learning
    Transactions of the Association for Computational Linguistics
    Hanjun Cho, Gahyun Yoo, Hanseong Kim, Jay-Yoon Lee

Awards

  • 2021
    EFL 8th Best Speaker, SNUDA X KNC 2021
    Korea Intervarsity Debate Association
  • 2014
    Bronze Medal, Korean Mathematical Olympiad
    Korean Mathematical Society

Presentations

  • Invited Seminar @ Kakao Mobility (Feb 2025)

ExtracurricularActivities

AttentionX: Member of Artificial Intelligence Research Group (2023)
Waffle Studio: Member of Seoul National University Web Development Club (2021-2023), developed the backend server of an interactive web editor application
SNUDA: Finance Minister and Member of Seoul National University Debate Club (2020-2021)

Skills

Languages: Python, C++, Java
ML Frameworks: PyTorch, SGLang, VLLM, Tensorflow, Scikit-Learn
Infrastructure: Docker

Coursework

Math: Linear Algebra, Calculus 1&2, Statistical Computing and Lab, Mathematical Analysis, Discrete Mathematics
Computer Science: Data Structures, Algorithms, Automata Theory, Logic Design, Computer Architecture, System Programming
Artificial Intelligence: Basics of Deep Learning, Machine Learning and Deep Learning for Data Science 1&2, Special Topic - Generative Modeling with Diffusion and Flow-Based Models