About Me

Masatoshi Nagahama (長濱 直智), Software Engineer at Sansan, Inc. ex-Tech Lead of Software/Data Engineer & Cloud Architect at CyberAgent, Inc.

数理最適化、機械学習（深層学習）、信号処理分野で修士号を取得。

Professional Experience

  • Software Engineer
  • Data Engineer
  • Cloud Architect
  • Research Scientist

Work Experience

2024.09 - Current: Sansan, Inc., Software Engineer.

2022.04 - 2024.08: CyberAgent, Inc., Software Engineer.

DSPの開発に従事。サーバーのパフォーマンスチューニング、データ基盤構築、クラウド最適化などを行う。クラウドサービス、サーバレスアーキテクチャ、クラウドのコスト最適化、キャッシュシステムなどに詳しい。新卒向けにデータエンジニアリング、SQLの講義を担当した。

I am deeply involved in the development of DSPs (Demand Side Platforms), comprehensive tools meticulously designed to streamline the process of advertisement delivery and ensure precise audience targeting, enabling our clients to achieve peak advertising performance. My work encompasses the construction of data infrastructures leveraging Snowflake, server implementations using the Go language, and the development of robust infrastructures powered by AWS.

To guarantee rapid server response and efficiency, I employ advanced techniques such as in-memory caching and make use of tools like Memcached. Furthermore, drawing on the inherent concurrency features of Go, I ensure our servers are not only fast but also optimized for peak performance.

Additionally, I pride myself on my expertise in optimizing and speeding up database systems, particularly with MySQL and DynamoDB. With a holistic approach to system development, I ensure that every component, from data storage to server execution, is honed for excellence.

2024.02 - 2025.07: ALGO ARTIS, Inc., Software Engineer (contract).

2020.10 - 2024.01: SUPWAT, Inc., Software Engineer / Data Scientist (contract).

2021.09 - 2023.03: Megagon Labs., Software Engineer (contract).

2020.08 - 2021.04: Forcia, Inc., Data Scientist (part-time): dynamic pricing algorithm and systems.

ダイナミックプライシングの予測モデルの開発に従事。Python (Machine Learning)、TypeScipt, PostgreSQL etc…

2019.10 - 2020.07: ****, Data Scientist (part-time).

自然言語処理を用いた推薦システムの構築に従事。

I was deeply involved in tasks pivotal to natural language processing (NLP). My responsibilities included web scraping for data collection tailored for language learning, utilizing MeCab for morphological analysis, and employing tf-idf for word ranking. These tasks were all instrumental in building a recommendation system that leveraged the power of NLP. My role not only allowed me to deepen my expertise in NLP but also to understand the practical application of these techniques in the construction of recommendation engines.

Engineering Skills

Languages & Frameworks

  • Go — メインで使用。並行処理、サーバー、CLI など
  • Python — FastAPI, Flask, PyTorch, gensim, Scrapy, Poetry など
  • TypeScript / JavaScript — React, Next, Vue, Nuxt, Chakra UI, MUI, Vuetify など

Architecture & Design

  • Domain-Driven Design (DDD)
  • Distributed Systems / Distributed Architecture

Data & Storage

  • RDB / Data Warehouse: MySQL, PostgreSQL, Snowflake, Cloud Spanner
  • NoSQL / Cache: DynamoDB, Redis, Memcached

Cloud & Infrastructure

  • AWS — 4 certified ⭐️
  • Google Cloud — Cloud Spanner, BigQuery, Cloud Run, Cloud Functions など
  • Snowflake, Datadog
  • Terraform (AWS, Snowflake), Docker, GitHub Actions

Performance & Tuning

  • サーバー・DB・SQL のチューニング、インメモリキャッシュ設計

Tools & Others

  • GitHub, RabbitMQ, zsh, bash — Mac / Linux

Certifications

Research Interests

  • Graph Signal Processing: Signal recovery, denoising, and restoration
  • Mathematical Optimization: convex optimization, optimization algorithm
  • Image Processing: Image coding and image compression
  • Machine Learning (Deep Learning)

📗 Publications

Journal Paper

  1. M. Nagahama, K. Yamada, Y. Tanaka, S. H. Chan, and Y. C. Eldar, “Graph signal restoration using nested deep algorithm unrolling,” IEEE Transactions on Signal Processing, vol. 70, pp. 3296-3311, 2022. [Project Page] [IEEE Xplore] [pdf]

Conference Paper

  1. M. Nagahama and Y. Tanaka, “Multimodal graph signal denoising via twofold graph smoothness regularization with deep algorithm unrolling,” IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022, Singapore, May 2022. [IEEE Xplore]
  2. M. Nagahama, K. Yamada, Y. Tanaka, S. H. Chan, and Y. C. Eldar, “Graph signal denoising using nested-structured deep algorithm unrolling,” IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2021, Virtual Conference, Jun. 2021. [IEEE Xplore]

Awards / Grants

  • 日本学生支援機構 大学院第一種奨学金 特に優れた業績による返還免除 (2020.4 - 2022.3)

📗 Books

SnowflakeとTerraformで作るデータ基盤入門

Building Modern Data Platforms: A Guide to Snowflake and Terraform

🗣️ Talks

  • 2023.12 ディビジョン内勉強会 「パフォーマンスチューニング・クラウドコスト最適化」 60min
    • DSPのパフォーマンスチューニングやクラウドのコストを下げるためのポイントを解説、キャッシュシステムなどの解説
  • 2023.12 チーム内勉強会 「Goを書く技術」 60min — Go言語上級向け
  • 2023.5 社内新卒研修 「データベース・SQL研修」 90min
  • 2023.5 社内新卒研修 「データエンジニアリング研修」 90min — データ基盤の作り方などを新卒向けに解説

Blog / Posts

はてなブログ 最近記事:

OSS / Software

SNS / Links