About

avatar

Minato Sato

Embedded AI Engineer
Sony
githubGitHublinkedinLinkedinxX

About Me

For over eight years, I’ve been dedicated to developing AI-centric products and services, building a robust skill set in software engineering, project management, small team leadership, and data science. Alongside my primary role, I also freelance, focusing on language and speech processing as well as recommendation systems powered by large language models (LLMs). My expertise spans both hands-on development and high-level consulting, partnering with companies of various sizes to deliver AI-driven solutions.

On this website, I write blog posts primarily about natural language processing, machine learning, and software engineering in Japanese.

GitHub / Twitter / Facebook / LinkedIn / Google Scholar

Skills

C++, C, RTOS, Python(+Cython), JavaScript(TypeScript), MySQL, Imapla, Java, R, C#, AWS/GCP, Soldering

Work Experience

AI Consultant, Research Engineer

Freelancer
Yokohama, Kanagawa Prefecture, Japan
February 2023 - Present

I work with companies on a contract basis to implement a wide range of AI solutions, including cutting-edge recommendation systems, retrieval-augmented generation (RAG) powered by large language models, and sophisticated speech-driven conversation systems. I offer comprehensive development and consulting services to help organizations seamlessly integrate these technologies into their products and workflows.

Embedded AI Engineer

Sony Corporation
Yokohama, Kanagawa Prefecture, Japan
April 2022 - Present

I’m currently focused on developing advanced subject-recognition AI for mirrorless digital cameras, enabling effortless image capture without the usual focus concerns. In addition to my core software engineering responsibilities, I refine system performance through field testing and in-depth interviews with industry professionals. Most recently, my work has expanded to include developing object tracking algorithms and creating hardware drivers for recognition AI, ensuring our technology remains robust and cutting-edge.

Products I have been involved in: α7RV, α9III, α1II

Machine Learning Engineer, Data Scientist

So-net Media Networks Co.,Ltd.
Shinagawa-Ward, Tokyo, Japan
October 2018 - June 2022

In my role as a senior machine learning engineer and data scientist, I led the end-to-end development and operation of AI-driven user targeting products for internet advertising. By building real-time user targeting systems and ensuring continuous, stable operations, I delivered significant sales gains that positively influenced key performance metrics.

Products I have been involved in: VALIS Engine

Software Engineer

Sony Imaging Products & Solutions Inc.
Atsugi, Kanagawa Prefecture, Japan
April 2017 - September 2018

I contributed to the development of an “AV over IP” solution built around CLED, a state-of-the-art large-scale LED display technology.

Data Scientist, Analyst (Part-time Job)

National Maritime Research Institute (国立研究開発法人 海上技術安全研究所)
Mitaka, Tokyo, Japan
November 2014 - March 2017

I analyzed ship operation monitoring data to evaluate and estimate a vessel’s propulsion performance. I also authored research papers documenting my analytical methods and findings.

Education

The University of Electro-Communications, Tokyo, Japan(2015-2017)

  • M.Eng in the Department of Social Intelligence and Informatics
    • The salutatorian, the runner-up in the department.
  • Research on Natural Language Processing using Deep Learning

The University of Electro-Communications, Tokyo, Japan(2011-2015)

  • B.Eng in the Department of Informatics
  • Research on Evolutionary Machine Learning (Reinforcement Learning)
  • Development Activities of Ultra-small Artificial Satellites (Exploration Rovers)

Publication

Google Scholar Link

Conference Papers (Reviewed)

  • Minato Sato, Ryohei Orihara, Yuichi Sei, Yasuyuki Tahara and Akihiko Ohsuga:Text Classification and Transfer Learning Based on Character-Level Deep Convolutional Neural Networks, in book "Agents and Artificial Intelligence" (Selected revised papers from ICAART 2017), ISBN: 978-3-319-93580-5, Springer, pp.62-81, 2018/06
  • Minato Sato, Ryohei Orihara, Yuichi Sei, Yasuyuki Tahara and Akihiko Ohsuga: Japanese Text Classification by Character-Level Deep ConvNets and Transfer Learning, 9th International Conference on Agents and Artificial Intelligence (ICAART 2017), Volume 2, pp.175-184, 2017/02 (acceptance rate: 26.5%, candidate to ICAART 2017 best student paper award)
  • Minato Sato, Kotaro Usui, Masaya Nakata and Keiki Takadama: Detecting Shoplifting From Customer Behavior Data by Extended XCS-SL: Towards Feature Extraction on Class-Imbalanced Sequence Data, IEEE Congress on Evolutionary Computation 2015 (CEC 2015), Sendai, Japan.

Regional Papers (Reviewed)

  • 佐藤挙斗，折原良平，清雄一，田原康之，大須賀昭彦: "文字レベル深層学習による日本語テキストの分類と転移学習"，合同エージェントワークショップ＆シンポジウム2016 (JAWS2016)，2016/09

Regional Papers (Unreviewed)

  • 佐藤挙斗, 折原良平, 清雄一, 田原康之, 大須賀昭彦: 文字レベル深層学習によるテキスト分類と転移学習, 人工知能学会 第102回 人工知能基本問題研究会(SIG-FPAI), pp.13-19, 2016/12
  • 佐藤挙斗，折原良平，清雄一，田原康之，大須賀昭彦: 文字レベル深層学習の日本語データセットへの応用，第184回 情報処理学会 知能システム研究会(SIG-ICS)，2016/08
  • 加納敏幸, 佐藤圭二, 佐藤挙斗:運航モニタリングデータによる船舶推進性能評価，日本船舶海洋工学会平成27年秋季講演会，2015/11
  • 小峯嵩裕，佐藤挙斗，田島友祐，中田雅也，高玉圭樹: 動的環境におけるアーカイヴ型学習分類子システム， 進化計算シンポジウム2013，pp.263-268，2013/12

Achievement