Skip to content
View Mephisto405's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report Mephisto405

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Mephisto405/README.md

In-Young Cho

Machine Learning Engineer
πŸ“ž (+82) 10-9504-8680 | πŸ“§ [email protected] | 🌐 linkedin.com/in/in-young-cho | πŸ“ Seoul, South Korea

About πŸš€

Machine Learning Engineer with over 3 years of professional experience, specializing in developing 2D/3D generative AIs to expand users' creative possibilities. Expertise in designing scalable algorithms to enhance AI workflows, working closely with research scientists, engineers, and product managers to prioritize solutions that maximize user impact. Skilled in Python, PyTorch, Blender, and experienced with cloud infrastructure like Azure and AWS. Passionate about turning ideas into outstanding user experiences and solving business challenges.

Experience πŸ’Ό

Deep Learning Researcher @ KRAFTON Inc.
πŸ“ Seoul, South Korea | Nov. 2022 - Present

  • Developed a scalable algorithm for collecting refined training data from large-scale 3D assets and worked with diffusion models using ControlNet and LoRA, enhancing the quality of multi-modal PBR texture generation.
  • Reduced mesh processing time from hours to seconds by designing an algorithm to convert unstructured meshes into watertight forms, reducing complexity from cubic to quadratic.
  • Optimized a 3D reconstruction model, generating game assets from a single photo in under 30 seconds on consumer devices. This was recognized as a top AI innovation for the upcoming flagship game, inZOI.

Machine Learning Engineer @ Spacewalk Inc.
πŸ“ Seoul, South Korea | Jul. 2021 - Sep. 2022

  • Designed and deployed Point Transformer-based reinforcement learning algorithms (A3C, PPO) for optimizing building designs, improving profit maximization while complying with regulations.
  • Applied pairwise ranking learning for reward function tuning, reducing tuning time by 200x and increasing agreement with architect preferences from 55% to 90%, doubling AI outputs meeting human-quality standards to 60%.

Education πŸŽ“

MSc in Computer Science @ KAIST
πŸ“ Daejeon, South Korea | Sep. 2019 - Aug. 2021

BSc in Computer Science and Mathematics (Double Major) @ KAIST
πŸ“ Daejeon, South Korea | Mar. 2015 - Aug. 2019

  • GPA: 4.00/4.30, summa cum laude
  • Honors: Dean's list (top 3%)
  • Relevant coursework: Computer graphics, operating systems, real/complex/numerical analysis, linear algebra, matrix computation, probability theory, statistics

Publications πŸ“š

  1. Rengan Xie, Kai Huang, In-Young Cho, Sen Yang, Wei Chen, Hujun Bao, Wenting Zheng, Rong Li, Yuchi Huo. ReN Human: Learning Relightable Neural Implicit Surfaces for Animatable Human Rendering. ACM Transactions on Graphics, 2024.
  2. In-Young Cho, Yuchi Huo, Sung-Eui Yoon. Weakly-Supervised Contrastive Learning in Path Manifold for Monte Carlo Image Reconstruction. ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH), 2021.

Patents πŸ”‘

  1. Sung-Eui Yoon, In-Young Cho, Yuchi Huo. Ray Clustering Learning Method Based on Weakly-Supervised Learning for Denoising Through Ray Tracing. US Patent 12051146B2, 2024.

Skills πŸ› οΈ

  • Languages: Korean (native), English (professional working proficiency, TOEIC: 910/990)
  • Programming Languages: Python, C/C++, CUDA
  • Software: Git/Hub/Lab, Linux, Windows, Blender, Jenkins, MkDocs
  • Libraries: PyTorch, scikit-learn, NumPy, Pandas, Blender Python API
  • Cloud/ML Platforms: Azure, AWS, MLflow, Kubeflow

Last updated: October 2024

Pinned Loading

  1. WCMC WCMC Public

    Weakly-Supervised Contrastive Learning in Path Manifold for Monte Carlo Image Reconstruction [Cho et al. SIGGRAPH 2021]

    Python 32 3

  2. Learning-Loss-for-Active-Learning Learning-Loss-for-Active-Learning Public

    Reproducing experimental results of LL4AL [Yoo et al. 2019 CVPR]

    Python 215 50

  3. Unsupervised-Out-of-Distribution-Detection-by-Maximum-Classifier-Discrepancy Unsupervised-Out-of-Distribution-Detection-by-Maximum-Classifier-Discrepancy Public

    Reproducing experimental results of OOD-by-MCD [Yu and Aizawa et al. ICCV 2019]

    Python 31 3

  4. NeRF-Reproduce NeRF-Reproduce Public

    CS492 3D-ML Term Project by In-Young Cho

    Python 3

  5. OptaGen OptaGen Public

    Optix-based automated data generation tool

    C 1

  6. Accelerated-Ray-Tracing-in-One-Weekend-in-CUDA Accelerated-Ray-Tracing-in-One-Weekend-in-CUDA Public

    Inspired by https://devblogs.nvidia.com/accelerated-ray-tracing-cuda/

    C++