Jincheol Yang

Ph.D. candidate in Electronic Engineering
Sogang University, Seoul, South Korea

I'm a Ph.D. candidate at Sogang University, advised by Prof. Suk-Ju Kang. I work on efficient AI for computer vision: making large vision models smaller and faster to run while keeping their accuracy.

My recent work focuses on post-training quantization (PTQ) for vision transformers (ViTs) and monocular depth estimation (MDE), and on KV cache compression and sparse attention for visual autoregressive (VAR) and video models. I also collaborate with Samsung Electronics and Samsung Display on model compression, energy-efficient hardware acceleration, and image restoration.

  • Post-training quantization
  • KV cache compression
  • Sparse attention
Portrait of Jincheol Yang

Publications

† Corresponding author

Research experience

Industry collaborations and internships

  1. Sep 2025 – Present

    Samsung Electronics

    Research collaboration

    Memory-efficiency-focused model compression and optimization techniques for large-scale deep learning models.

  2. Mar 2025 – Present

    Samsung Display

    Research collaboration with the Computer-Aided Engineering team

    Developing AI technology for image quality restoration.

  3. Sep 2024 – Sep 2025

    Samsung Electronics

    Research collaboration

    Developed a memory-fused, transfer-operation-based deep learning hardware accelerator for better energy efficiency.

  4. Jul 2023 – Feb 2025

    Samsung Display

    Research collaboration with the Computer-Aided Engineering team

    Developed AI technology for image-based depth estimation.

  5. Jul 2022 – May 2023

    Korea Electronics Technology Institute (KETI)

    Research intern, Intelligent Image Processing Research Center

Awards

  1. Aug 2026

    Best Poster Award

    The 26th International Meeting on Information Display (IMID 2026)

  2. Dec 2025

    Best Paper Award

    Department of Electronic Engineering, Sogang University

  3. Dec 2025

    Outstanding Paper Award

    Sogang University