cv

My curriculum vitae, keeping brief for anonymity.

Basics

Nickname masuwaka
Label An engineer
Email masuwaka@masuwaka.net
Url https://masuwaka.net/
Summary A Japanese AI researcher, data scientist, and system administrator of AI computing clusters.

Work

  • 2024.10 - Present
    System administrator of AI computing clusters
    System administration of large AI computing clusters (w/ hundreds of nVidia GPUs).
    • LDAP
    • Ansible
    • Zabbix
    • CUDA
    • Linux (Ubuntu)
  • 2017.10 - Present
    AI research engineer / data scientist
    Research and development of state-of-the-art AI technologies. I am also working on data mining from big data.
    • Neural Network
    • Bayesian Optimization
    • Tensorflow (Keras)
    • PyTorch
    • Optuna
    • Pandas
    • Python
    • Linux (Ubuntu)
  • 2016.04 - 2017.09
    ECC (Error Correction Codes) engineer
    Engaged in research on energy-efficient, reliable ECCs. I also developed Python environment for ECC analysis.
    • LDPC Codes
    • BCH Codes
    • C/C++
    • Python
    • Linux (RHEL)

Education

  • 2011.04 - 2016.03

    Yokohama, Japan

    PhD
    A Japanese National University
    Engineering
    • Wireless Communications

Skills

Mathematics
Probability
Statistics
Coding Theory
Communication Theory
Signal Processing
Information Theory
Programming
C/C++
Python
Shell script (bash/zsh)
PyTorch
Numpy
Pandas
Optuna
gPyTorch
BoTorch
System Management
LDAP
Ansible
Zabbix
Linux

Languages

Japanese
Native speaker
English
Conversational

Interests

Mathematics
Optimization Theory
Manifold Theory
Quantum Information Theory
Quantum Computation
Programming
Rust
Polars
cuDF
cuPy
SIMD
Omniverse

Projects

  • 2025.07 - Present
    A fast and highly interpretable Python library for manipulating multivariate Gaussian distributions
    Manipulations of Gaussian distributions are sometimes too cumbersome and troublesome especially when the distributions are multivariate. In this project, I am developing a Python library that enables calculation of product, quotient, marginal, and conditional distributions of multivariate Gaussian distribution more easily and with higher interpretability.
    • Multivariate Gaussian Distributions