ksy9164/resume

Seongyoung-Kang's Resume

Seongyoung Kang

---

Jeongneung-ro 8-gil, Seongbuk-gu, Seoul, Republic of Korea
[email protected] | Github : https://github.com/ksy9164
([phone] (KR) | Linked-in : https://www.linkedin.com/in/seongyoung-kang/

---

Education

University of Kookmin University, B.S. Computer Science Seoul, South Korea
* Major GPA: 3.86/4.5
* Key course in Computer Science : Operating System,Computer Archtecture,Data structure,Heterogeneous computing, Big Data platform, Software architecture.

---

Experience

EV-Lock

LINE Blockchain Competition (1st Prize)

SmartMeter Developer

• Developed SmartMeter for P2P energy trading that improved Device security.
        Implemented secure store platform by using ARM TrustZone.
        Implemented a platform that supports using SmartContracts on SmartMeter.

---

Lablup Inc.

Winter Intership

Backend.AI Package Maintainer

Dec,18-Feb,19

• Implemented Debian Packages, Debian Repository.
        I have Packaged all the dependent modules needed for Backend.AI installation.
        Implemented Debian Repository for release Backend.AI Packages.
        Made Backend.AI installation script.

---

University of California, Irvine

Dynamic Resource Management for Mobile Heterogeneous Platform

Advisor : Prof. Nikil Dutt

Profiler, Benchmark Developer

Jun,18-Aug,18

• Implemented CPU,GPU Dynamic Profiler for SnapDragon-835 platform.
        The CPU Profiler was implemented by running Perf for each CPU core,
        The GPU Profiler was implemented by reading information from the kernel.
         I also have implemented a multithreaded program to do this in parallel.
• Implemented benchmarks
         Modified the Deep Learning Code(Caffe) for mobile platform.
         Implemented Bash shell scripts to automate the benchmark and Python code for make graph.
         Modified canny edge detector code for CPU(OpenMP), GPU(OpenCL).

---

Kookmin University

KTUNE, The Deep Learning API for Manycore CPU

Advisor : Prof. Eun-jin Im

Researcher

Dec,16-Dec,17

• Optimized Deep Learning functions for Manycore CPU.
        Multithreaded Deep Learning code (OpenMP).
        Adapted MKL(Math Kernel Library) to Deep Learning code .
        Vectorized Deep Learning code (AVX-512).
• Ported Deep Learning functions to Python.
        Implemented wrapper(Boost Python) to porting C code to Python

---

Seoul National University

Accelerator Programming Summer School

Advisor : Prof. Jae-jin Lee

Finished course of GPGPU (OpenCL, CUDA)

Aug,17

---

Honors & Awards

• 2018 LINE blockchain competition - 1st prize
• 2017 Chairman of Korea science foundation -1st prize
• 2017 CPS creativity competition Daegu mayor prize - 3rd prize

---

Papers

Korean Institute of Information Scientists and Engineers

• Seongyoung-Kang and Eunjin-Im “Ktune : Autotuner for Parallelizing Backpropagation”. 2017.12

---

Skills

• Computer language : C, C++, Bash, Python, NodeJS, Solidity, CUDA, java.
• Platform / Tools : Ubuntu, AWS, Docker, Git, Hadoop, Snapdragon, Samsung Artik.
• Language: Korean, English(Basic).