Avatar

Junyeong Ahn

AI Researcher

@{HYU, KAIST AI}



I'm an incoming M.S./Ph.D. integrated student at KAIST GSAI. Currently, I'm an undergraduate student in the Dept. of Data Science at Hanyang University.

I am also a research intern at the DAVIAN Lab, KAIST AI. Previously, I worked as an AI researcher at dob Studio Inc., and as a research intern at the Samsung Electronics DS AI Center.

My research interests include 3D vision and diffusion models.

Contact

Education

  • BS in Data Science, 2020 - 2026 (Expected)

    Hanyang University, Korea

    Major GPA: 4.41 / 4.5 (98.6%)

  • MS/Ph.D Integrated in AI, 2026 (Confirmed) -

    KAIST, Korea

    Advisor: Prof. Jaegul Choo

Industry Experiences

  • Samsung Electronics, Gyeonggi, South Korea
    Mar. 2025 - Jun. 2025

    Research Intern, AI Center.
  • dob Studio Inc., Seoul, South Korea
    Sep. 2023 - Oct. 2024

    AI Researcher, R&D Center.

Research Experiences

  • KAIST, Seongnam, South Korea
    Sep. 2024 - Present

    Research Intern, DAVIAN Lab.
  • Depaul University, Chicago, USA(Remote)
    Nov. 2022 - Sep. 2024

    Research Intern, Data Science Innovation Lab.
  • Seoul National University, Seoul, South Korea
    Dec. 2023 - Feb. 2024

    Explore CSR Student, SKI-ML Lab.
  • Hanyang University, Seoul, South Korea
    Oct. 2020 - Feb. 2021

    Research Intern, Data Science Innovation Lab.

Publications

  • TexAvatars: Hybrid Texel-3D Representations for Stable Rigging of Photorealistic Gaussian Head Avatars

    *Junyeong Ahn, *Jaeseong Lee, Taewoong Kang, Jaegul Choo

    3DV 2026, Vancouver, Canada

  • Blind Review (Concept Erasure Method)

    *Junyeong Ahn, *Seojin Yoon, Sungyong Baik

    Under review: CVPR 2026

  • Digital Health Sensor Data in Autism: Developing Few Shot Learning Approaches for Traditional Machine Learning Classifiers

    Casey C Bennett, Cedomir Stanojevic, Jiyeong Oh, Junyeong Ahn, Yeeun Jeon, Kanghee Son, Nikki Abbott

    IEEE BSN 2024, Chicago, USA

    [ paper ]

Projects

  • Draw Your Day


    Efficient AI-enhanced Diary-to-Illustration Experience, using LLaMa-2 with RAG as a promptiser

    [ project link ]

  • NAT-Diffuser


    Discrete Absorbing Diffusion Meets Neighborhood Attention Transformers at Vector-Quantized Space

    [ paper ] [ project link ]