YoonseopKang

blog

I am working to make computers see, hear and think.

Personal Interest

Artificial Intelligence, deep learning, recommendation systems, multimodal learning, computer vision.

Employment

  1. Feb 2018–Present CTO & Co-founder of SIGTERM Inc.
    Developed high-frequency trading (HFT) system, operating on multiple major crypto markets.
    Coordinated a team of 3~5 software engineers: established a code review and evaluation process.
    Established guidelines and policies for code quality improvement (e.g. code review process, architecture reviews, automated linting and testing).
    Developed low-latency order execution and management software for the HFT system.
    Co-worked on programming language transition of the HFT system core.
    Developed highly optimized crypto exchange clients for the HFT system.
    Improved speed and stability of orderbook retrieval and parsing component.
  2. Feb 2016–Feb 2018 Lead Data Scientist, WATCHA Inc.
    Implemented recommendation engine for WATCHA! for books.
    Improved profitability of WATCHAPLAY by optimizing recommendations.
    Distributed and optimized data processing pipeline.
    Developed recommendation algorithms for home/search/category page of WATCHAPLAY.
    Coordinated a research team of 3–5 people.
  3. Mar 2014–Feb 2016 Data Scientist, WATCHA Inc.
    Improved recommendation accuracy and optimized the sign-up procedures of WATCHA!.
    Developed content recommendation algorithm for home feed of mobile version of WATCHA!.
    Modeled relationships between user preferences on contents and other attributes for use in targeted advertisement.
    Contributed porting python-based recommendation system to Scala.

Education

  1. Feb 2008–Feb 2014 Ph. D. student of Machine Learning Group @ POSTECH.
    Presented 7 peer-reviewed papers about machine learning.
    Participated in 2 national, > 2 coorperate research projects.
  2. Mar 2004–Feb 2008 Computer Science, POSTECH, B.S.
  3. Mar 2002–Feb 2004 Gyeonggi Science High School

Publications

International Conferences

  1. Yoonseop Kang, Taewoong Jang, Seungjin Choi, Learning features with Structure-adapting multi-view exponential family harmoniums, IJCNN, July 6–11, 2014.
  2. Yoonseop Kang, Saehoon Kim, Seungjin Choi, Deep Learning to Hash with Multiple Representations, ICDM, Dec 10–13, 2012.
  3. Saehoon Kim, Yoonseop Kang, Seungjin Choi, Sequential Spectral Learning to Hash with Multiple Representations, ECCV, Oct 7–13, 2012.
  4. Yoonseop Kang, Seungjin Choi, Restricted Deep Belief Networks for Multi-View Learning, ECML-PKDD, Sep 5–9, 2011.
  5. Heeyoul Choi, Seungjin Choi, Anup Katake, Yoonseop Kang and Yoonsuck Choe, Manifold Alpha-Integration, PRICAI, Aug 20–September 2, 2010.
  6. Yoonseop Kang and Seungjin Choi, Common neighborhood sub-graph density as a similarity measure for community detection, ICONIP, Dec 1–5, 2009.
  7. Heeyoul Choi, Anup Katake, Seungjin Choi, Yoonseop Kang, and Yoonsuck Choe, Probabilistic combination of multiple evidence, ICONIP, Dec 1–5, 2009.

Softwares

e0en's contributions

See more at https://github.com/e0en.

Projects

  1. Mar 2018–Now A high-frequency trading system for SIGTERM Inc.
  2. Mar 2014–Feb 2018 Recommendation system for WATCHAPLAY—a subscription VOD service
  3. Mar 2014–Feb 2018 Recommendation system for WATCHA!—a Movie/TV show recommendation service
  4. Jul 2009–Feb 2014 Development of models for feature extraction from multi-sensory data
    Supported by Ministry of Education and Science Technology
  5. Sep 2008–Dec 2009 Project Collaboration: CMEST—Vision interfaces for mobile devices
    Supported by Ministry of Commerce, Industry and Energy
  6. Sep 2008–Dec 2008 Teaching assistant: Automata and Formal Languages.
  7. Mar 2008–Jun 2008 Teaching assistant: Introduction to Computer Science.
  8. Sep 2007–Dec 2007 Senior Project: Automatic Transcription of Music (advisor: Seungjin Choi).

Skills

Programming Languages

Library/Frameworks

Other tools