Kohei Sugihara

Tsukuba, Japan

I am a software engineer working for Preferred Networks, Inc., a deep-learning startup company in Tokyo, Japan. My areas are planning, designing, and operating in-house storage clusters of several object storage systems (> 10 PB in logical) from physical (i.e. data center facility management) to logical layers (i.e. performance monitoring/optimization, user supports, system integration with a company-wide ML/DL platform on on-premise Kubernetes).

I am also a doctoral student in Computer Science and a student of High-Performance Computing System (HPCS) laboratory in University of Tsukuba. My research topic is distributed storage systems and system softwares in High-Performance Computing (HPC).

Other Resources

github.com/k5342 www.linkedin.com/in/sugiharak View my public profile

Computer Skills and Interests

  • Distributed Systems, High-Performance Computing (HPC), Parallel File Systems, Distributed Storage
  • C, Golang, Java, Ruby, Python
  • Linux, Apache Ozone, Apache Hadoop

Work Experiences

  • Full-time Engineer, Preferred Networks, Inc., 2021/04 - present
    Storage Engineer

Education

Professional Training

  • Part-time Engineer, Preferred Networks, Inc., 2018/11 - 2019/03
  • Summer Internship, Preferred Networks, Inc., 2018/08 - 2018/09
    Research and development activities on a high-throuput RDMA protocols and distributed cache system for training data in the distributed deep learning using native InfiniBand Verbs.
  • Summer Internship, Wantedly, Inc., 2017/08 - 2018/09
    I worked as an summer internship project with the mobile application team to investigating user retention data collected from the application usages. I’ve spent a days that writing 30+ lines of SQL during the internship. My contribution is providing a new user statistics dashboard on DOMO including some bug fixes in the logging platform written in Rails ans Golang.
  • Part-time Engineer, Hematite, LLC., 2017/04 - 2020/10
  • Part-time Engineer, Stellalink, Inc., 2014/08 - 2017/03
    Web system developments in Java, PHP and MySQL.

Publications

Peer-reviewed Conference Papers

  • Kohei Sugihara, Osamu Tatebe, “Design of Direct Read from Sparse Segments in MPI-IO”, Proceedings of the 6th IEEE International Conference on Data Science and Systems (DSS), pp.1308-1315, doi: 10.1109/HPCC-SmartCity-DSS50907.2020.00168, 2020. [Acceptance Rate: 14/69 = 20%]
  • Kohei Sugihara, Osamu Tatebe, “Design of Locality-aware MPI-IO for Scalable Shared File Write Performance”, Proceedings of the IEEE International Workshop on High-Performance Storage, pp.1080-1089, doi: 10.1109/IPDPSW50202.2020.00179, 2020.

Others

Other Activities

  • Security Camp 2013 (as a student in the Web Security class)
  • Security Camp 2018 (as a tutor in the Junior Security class)
  • ISUCON 7 Finalist (in student quota), 2017
  • ISUCON 10 Finalist (in student quota), 2020