Jaesung Park

[email protected] | Github | CV

Senior Software Engineer (Ph.D.) specializing in GPU systems, real-time rendering, and performance-critical C++ infrastructure.
Architect and optimize GPU-accelerated systems for 3D rendering and ML workloads using Vulkan and CUDA.
Former IOI Gold Medalist and ICPC World Finalist.

Work Experience

• Presto Labs -- Quant Researcher Sep 2023 -- Present

  • Designed and maintained low-latency C++ infrastructure for research and production deployment.
  • Improved simulation throughput by up to 10x compared to legacy systems.

• NAVER LABS -- Research Software Engineer Feb 2022 -- Sep 2023

  • Implemented high-performance rendering pipelines for massive point clouds in OpenGL applications.
  • Designed Vulkan-based real-time Neural Radiance Fields (NeRF) rendering engine with GPU shader pipeline.
  • Implemented WebGL-based 3D visualization for real estate virtual tours (contributed to KR patent).

• Cupix -- Research Software Engineer Jul 2020 -- Feb 2022

  • Developed compression algorithms for large-scale unstructured point clouds.
  • Implemented real-time WebGL rendering systems for interactive 3D visualization in browser environments.
  • Contributed to indoor 360° panorama reconstruction and photogrammetry systems.

• Moloco -- Software Engineer Intern May -- Aug 2017

  • Performed data analysis.
  • Contributed to data infrastructure engineering.

Programming Skills

• Languages: C++17/20, Python, JavaScript, TypeScript
• GPU / Graphics: CUDA, Vulkan, OpenGL, WebGPU, WebGL
• Systems: Performance optimization, low-latency systems, parallel computing
• ML / Vision: Neural Radiance Fields, Gaussian Splatting

Education

• Ph.D. in Compute Science Sep 2015 -- May 2020
  University of North Carolina at Chapel Hill USA

  • Advisor: Prof. Dinesh Manocha
  • Research: Robot motion planning, collision detection, ML-based human motion prediction.

• B.S. in Computer Science, Minor in Mathematics Mar 2011 -- Feb 2015
  Seoul National University South Korea

  • GPA: 4.06/4.30 (cumulative), 4.22/4.30 (major), Summa Cum Laude

Research Experience

• Optimization-Based Robot Motion Planning 2015 -- 2020

  • Jae Sung Park, Chonhyon Park, Dinesh Manocha.
    I-Planner: Intention-Aware Motion Planning Using Learning-Based Human Motion Prediction.
    The International Journal of Robotics Research (IJRR) 38 (1), 23-39, 2019.
  • Jae Sung Park, Chonhyon Park, Dinesh Manocha.
    Intention-Aware Motion Planning Using Learning Based Human Motion Prediction.
    Robotics: Science and Systems (RSS), 2017.
  • Chonhyon Park, Jae Sung Park, Steve Tonneau, Nicolas Mansard, Franck Multon, Julien Pettre, Dinesh Manocha.
    Dynamically Balanced and Plausible Trajectory Planning for Human-Like Characters.
    Proceedings of the 20th ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games. ACM, 2016.

• Learning-Based Human Intention Prediction 2016 -- 2020

  • Jae Sung Park, Dinesh Manocha.
    HMPO: Human Motion Prediction in Occluded Environments for Safe Motion Planning.
    Robotics: Science and Systems (RSS), 2020.
  • Jae Sung Park, Biao Jia, Mohit Bansal, Dinesh Manocha.
    Generating Realtime Motion Plans from Attribute-Based Natural Language Instructions Using Dynamic Constraint Mapping.
    IEEE International Conference on Robotics and Automation (ICRA), 2019.
  • Jae Sung Park, Chonhyon Park, Dinesh Manocha.
    Human Motion Prediction from Noisy Point Cloud Data for Human-Robot Interaction.
    IEEE RO-MAN workshop on Communicating Intentions in Human-Robot Interaction, 2016.

• Probabilistic Collision Detection Under Uncertainty 2016 -- 2020

  • Jae Sung Park, Dinesh Manocha.
    Efficient Probabilistic Collision Detection for Non-Gaussian Noise Distributions.
    IEEE Robotics and Automation Letters 5.2 (2020): 1024-1031.
  • Chonhyon Park, Jae Sung Park, Dinesh Manocha.
    Fast and Bounded Probabilistic Collision Detection for High-DOF Trajectory Planning in Dynamic Environments.
    IEEE Transactions on Automation Science and Engineering (TASE) 15 (3), 980-991, 2018.
  • Jae Sung Park, Chonhyon Park, Dinesh Manocha.
    Efficient Probabilistic Collision Detection for Non-Convex Shapes.
    IEEE International Conference on Robotics and Automation (ICRA), 1944-1951, 2017.
  • Jae Sung Park, Chonhyon Park, Dinesh Manocha.
    Fast and Bounded Probabilistic Collision Detection for High-DOF Robots in Dynamic Environments.
    Workshop on Algorithmic Foundations of Robotics (WAFR), 2016.

• Sweep-Based Surface Modeling 2013 -- 2014

  • Jaesung Park, Minsub Shim, Seon-Young Park, Yunku Kang, Myung-Soo Kim.
    Realistic deformation of 3D human blood vessels.
    Computer Animation and Virtual Worlds 24.3-4 (2013): 317-325.
  • Seon-Young Park, Jaesung Park, Minsub Shim, Yunku Kang, Myung-Soo Kim.
    Sweep-based Compression and Deformation of 3D Blood Vessel Models.
    HCI 2013 (2013): 11-14.

Awards

• ACM-ICPC World Finals: 36th place (2012), 51st place (2015)
• ACM-ICPC Daejeon Regional: 1st place (2011, 2014), 3rd place (2012), 5th place (2013)
• International Olympiad in Informatics (IOI): Gold Medal (2009), Silver Medal (2008)
• Asia Pacific Informatics Olympiad (APIO): 1st place (2009)
• Korea Computer Graphics Society Thesis Competition: Excellence Prize (2014)
• Korean Mathematical Competition for University Students, Major Division: Bronze Prize (2013, 2014)
• Undergraduate Research Program (URP): First Prize (2012)

Scholarship/Assistantship

• Doctoral Merit Assistantship, University of North Carolina at Chapel Hill, 2015
• National Science and Engineering Undergraduate Scholarship, South Korea, 2011-2014

Personal Projects

• vulkan_radix_sort: High-performance Vulkan-based GPU radix sort implementation.

  • Achieved performance competitive with CUDA CUB library.

• vkgs: Vulkan-based Gaussian Splatting viewer optimized for real-time performance.

  • Achieved 2× speedup over the original viewer through GPU pipeline restructuring and memory optimization.
  • Cited by Meta's vkraygs research project and referenced by NVIDIA’s Vulkan demo repository vk_gaussian_splatting.

• splatstream: Vulkan-based Gaussian Splatting viewer with Python bindings for research.