🔭 I’m currently working on Hyundai Autoever as a professional researcher

• Professinal Researcher in Hyundai Autoever (2021 ~ )
• B.S. degree in Computer Science
  • Kookmin University, Seoul, Korea (2014 ~ 2021)
• Student ID: 20143109
• Name: Jiwook Choi (최지욱, Woogie)
• GitHub ID: kkoma2623
• LinkedIn: Jiwook Choi
• BOJ(Baekjoon Online Judge) ID: boo106
• Blog: Velog
• email: [email protected]

• Professional Researcher
  • At: Hyundai Autoever
  • Term: May 2021 - Now
  • Location: Samseong-dong, Seoul, South Korea
  • What I'm doing:
    • Classic AUTOSAR
    • Develop module integrate automation tool
• Full Stack Engineer (Internship)
  • At: RiderDash
  • Term: Jun 2020 - Feb 2021
  • Location: Samseong-dong, Seoul, South Korea
  • What I've done:
    • Developed web
    • A/B and Multivariate testing
    • Improve conversion
  • What I did:
    • Web development: Used TypeScript, React.js, MobX, Redux.js, Apollo + GraphQL, NodeJS, Next.js, Nest.js
    • Server migration: Migrate GCP to AWS. Used S3, Lamda, Route53, CloudFront, IAM, SAM CLI
• Capstone Design Project
  • At: Kookmin University
  • Term: Mar 2020 - Jun 2020
  • Location: Seongbuk-gu, Seoul, South Korea
  • Award:
    • Popularity Award
  • Team: ㅇㅅㅇ (Algorithm Service Labatory, 알고리즘 서비스 연구소)
    • Myung-Seo Choi(PM), Jiwook Choi(FE Leader, me), Hyunwoong Woo(BE Leader), Hojoon Park, Sooryun Kang, Khalid
  • Whar I did:
    • Administrate FE develop schedule as a PL.
    • Use React and React hooks for FE develop
    • Use Phaser 3 game framework to implement replay page.
• Research Assistant(Online)
  • At: UC Irvine
  • Term: Dec 2020 - Feb 2021
  • Location: Onine
  • What I did: Studied DNN(Deep Neural Network), ZFP compression, Try to boost DNN performace by zfp compression algorithm on VGG.
• Dodam Dodam Project(Porn Image Filtering)
  • At: Kookmin University
  • Term: Nov 2019 - Dec 2019
  • Location: Seongbuk-gu, Seoul, South Korea
  • Team: Jiwook Choi(me), Hojoon Park, Hyobin Hwang, Hyunki Son, Yerin Oh
  • What I did:
    • Collect ans label 65,000 porn images
    • Learning pretrained ResNet50 and GoogLeNet
      • ResNet50's test set accuracy: 97%, recall: 95%, precision: 98%
      • GooLeNet's test set accuracy: 95%), recall: 91%, precision: 99%
    • Use OpenCV to detect image area and judege porn or not at realtime.
• UROP (Undergraduate Researcher)
  • At: Kookmin University
  • Term: Mar 2019 - July 2019
  • Location: Seongbuk-gu, Seoul, South Korea
  • What I did:
• FOSCAR
  • At: Kookmin University Autonomous Club
  • Term: Mar 2019 - Feb 2020
  • Location: Seongbuk-gu, Seoul, South Korea
  • Award:
    • 3rd Place at BRKL(Battle Robot Korea League) Autonomous Driving
    • Advanced to the finals as a representative of the Korean team at 2020 Turtlebot autorace
  • Team: Jiwook Choi(Leader, General Affairs, me), Myung-Seo Choi, Somang Kim, Kihoon Hong
  • What I did:
    • Use SVM and Hog features to detect traffic signs.
    • Use OpenCV for linetracing and traffic signs detect.
    • Use ROS to build robot applicatinos.
    • Use LDS sensor to detect obstacles.
    • Set parking algorithm.
    • Administrate club grants.
• KOBOT
  • At: Kookmin University Humanoid and Embedded Club
  • Term: Sep 2014 - Dec 2015
  • Location: Seongbuk-gu, Seoul, South Korea
  • Award:
    • Obstacle Run 2nd Place at FIRA RoboWorld Cup 2015
  • Team: Dakyung Oh(Leader), Hyungmin In, Jiwook Choi(me), Hyowon Choi
  • What I did:
    • Use OpenCV to detect obstacles.
    • Use RoboC to set humanoid robot's actions.
    • Get club grants more than 8,000,000 Won as a general affairs from Samsung Friendship and KOFST(The Korean Federation of Science and Technology).