Hiroshi Wayama

Software Engineer. Born in 1983 in Morioka, Iwate, Japan. Graduated from the Faculty of Science at Kyoto University and completed a master’s degree at the Graduate School of Science, Kyoto University, majoring in Physics.

In 2009, I joined Canon Inc., where I worked on semiconductor process development, numerical computation, simulation technologies, and web application development using AWS. While working as a full-time employee, I was also involved in image recognition projects using deep learning and various data analysis tasks. I left the company in 2019 and became independent.

In 2020, I co-founded LegalWin, a case management application for lawyers. In 2021, I sold a product that I had been developing since my time as a company employee. Currently, I work as a project manager and engineer on machine learning projects, while also conducting research on recommender systems as a working Ph.D. student. I live in Kyoto with my wife and two sons.

Contact

For inquiries, please use this form .

Entrepreneurship

1. LegalWin

A project co-founded with a lawyer friend. I am responsible for the technical leadership and am involved in daily development and operations.

To ensure reliability and usability for legal professionals, I developed a wide range of features required for the service, including frontend and backend systems, infrastructure, security, and the foundations for machine learning models and MLOps.

The service is currently in operation and is used by many lawyers. Lawyers and paralegals are very welcome to register.

2. Educational Web Application

A project I launched and developed independently as a sole proprietor while working as a company employee. It is a web application designed to help students study more efficiently.

The service was sold to a company in 2021 and is now operated under a different name, so no links are provided here.

3. Web Application Introducing Tourist Attractions in Kyoto

A web application I built independently while I was a university student. It introduced tourist attractions in Kyoto and was developed using HTML, CSS, JavaScript, PHP, and MySQL. Although it can hardly be called a startup, it was the first web application I built entirely on my own while learning from Linux server textbooks.

The site generated a small amount of advertising revenue, allowing me to earn money for the first time, which makes it a very special project to me.

Major Projects

A list of major projects I have been involved in as a sole proprietor or as part of companies I founded.

(* Projects undertaken as a company employee are excluded.)

1. LegalWin (Chief Technology Officer)

The case management application for lawyers mentioned above. Development started in 2018, a beta version was released in January 2020, and the official service launched in May 2020.

One of the co-founders, Attorney Hashimoto, implemented the frontend using Ruby on Rails and continues to improve the service daily to enhance efficiency for lawyers and paralegals.

As CTO, I was responsible not only for programming but also for infrastructure and database design to ensure high security and stable server operation at low cost. I also developed machine learning models and platforms to enable OCR and cloud-based document management features. Currently, I focus mainly on server operation and monitoring, in addition to feature updates, to ensure a stress-free user experience.

2. Educational Web Application Development (Business Owner)

An educational web application that I began developing independently in 2014.

I handled all aspects of development alone, including requirements definition, frontend and backend implementation, database design, and infrastructure construction. Since I was working full-time, development progressed gradually during evenings and weekends.

This was the first web application through which I earned income as an independent contractor rather than receiving a salary, making it a particularly memorable project.

3. Real-Time Recommendation System for an E-Commerce Site (Project Manager)

As a sole proprietor, I was contracted by a company to lead the development of a recommendation engine for an e-commerce site, from algorithm design to infrastructure construction.

As both Project Manager and Machine Learning Engineer, I was involved in technology selection, requirements definition, budget and schedule management, client communication, infrastructure design, and model development. The team members were highly skilled, and I learned a great deal from the experience. The warm words I received at the end of the contract are unforgettable.

4. Recommendation Algorithm Development for a Video Platform (Machine Learning Engineer)

I worked on developing the recommendation engine for a video platform as a contracted machine learning engineer. I proposed and implemented various algorithms based on collaborative filtering and content-based filtering.

I was involved with this company for a long period, from development to maintenance and operations, and am deeply grateful to the talented in-house engineers I worked with.

5. Recommendation Engine for Financial Products at a Financial Institution (Machine Learning Engineer)

I developed algorithms to recommend optimal financial products to individual customers. This project was contracted from the same company as the one described in Project 4.

6. Optimization Algorithm for Manufacturing Production Lines (Data Scientist)

I developed optimization algorithms to reduce costs and labor in a manufacturing production line. Although the project duration was short (about two weeks), I researched and implemented an optimal algorithm from scratch. It was a valuable opportunity to deepen my understanding of optimization. This project was also conducted as a contract.

7. Web Application Development Using LLMs (Web Application Engineer)

At a SaaS company, I worked on developing web applications utilizing large language models (LLMs). In an agile development environment typical of a startup, I was responsible for API design and implementation, as well as business logic incorporating AI algorithms.

Other Small-Scale Projects

1. Cheat User Detection Algorithm for Games (Data Scientist)

I worked from the PoC stage on developing algorithms to detect users violating mobile game terms of service (e.g., repeated cheating behavior). This was also a contracted project.

At the request of a university professor, I participated in a project analyzing the relationship between urban population trends and densely inhabited districts, where I was responsible for data analysis.

Publications

Peer-reviewed papers on recommender systems written as a working Ph.D. student.

  • Hiroshi Wayama and Kazunari Sugiyama: Generalized Personalized PageRank with Graph Convolutional Networks in Recommender Systems, The 15th International Conference on the Theory of Information Retrieval (ICTIR 2025), pp. 380–389, Padua, Italy, July 18, 2025. pdf (open access)

  • Hiroshi Wayama and Kazunari Sugiyama: The Effectiveness of Quantum Random Walk Model in Recommender Systems, The 13th International Conference on the Theory of Information Retrieval (ICTIR 2023), pp. 225–234, Hybrid (Taipei, Taiwan and Online), July 23, 2023. pdf (open access)

  • Hiroshi Wayama and Kazunari Sugiyama: A Recommender System Using Quantum Models, Journal of the Database Society of Japan, Vol.21-J, Article No.3, March 2023. pdf

Patents

A list of patents related to semiconductors and systems that I authored during my time as a company employee. These patents were written as part of my professional duties. The link below lists over 100 documents including co-authored works, though I was the primary inventor on about a dozen of them.

I also received a Regional Invention Encouragement Award (Kanto Region) as a co-inventor. While the primary inventor was exceptionally talented and I was listed as a co-inventor, it remains a great honor.

Skills

  • System Development
    • AWS
    • GCP
  • Machine Learning Model Development
    • Recommender Systems
    • Information Retrieval
  • Numerical Computation
    • Finite Element Method
    • Physical Simulation

Tools

  • PyTorch
  • SciPy
  • NumPy
  • Pandas
  • React
  • scikit-learn

Programming Languages

  • Python
  • Go
  • TypeScript
  • C/C++

Experience

  • 2020 –
    • Co-founder of LegalWin, Machine Learning Engineer
  • 2015 – 2019
    • Software Engineer at a manufacturing company
  • 2009 – 2015
    • Semiconductor Engineer at a manufacturing company

Education

  • 2021 –
    • Graduate School of Informatics, Kyoto University (Working Ph.D. Program, Recommender Systems)
  • 2007 – 2009
    • Graduate School of Science, Kyoto University (Quantum Optics Laboratory)
  • 2003 – 2007
    • Faculty of Science, Kyoto University (Physics)

Favorite Quotes

  • 人間万事塞翁が馬 (A blessing in disguise)
  • It is better to act and repent than not to act and regret.

Skill Map

This is a visualization of my professional experience.

After graduating from university, I started my career as a semiconductor development engineer at a manufacturing company. Later, I transferred to a systems development department and worked as a software engineer, becoming independent in 2019.

Based on my experience, effort is not always rewarded. However, I believe that without the effort I made in the past, it would have been impossible to live independently. I am satisfied with my current work in terms of fulfillment, enjoyment, and income, and in that sense, effort may indeed pay off.

I still have about 40 years ahead of me in life, and I look forward to seeing how things will change. This section is updated approximately once a year.

skill map

2003

During high school and immediately after graduation, I spent most of my time playing mahjong. There was even a period when I seriously considered making a living from it. I remember playing mahjong continuously for 24 hours over the New Year holidays, or spending Christmas Eve at a mahjong parlor—only to step outside after an all-night session and find snow falling, joking with friends about having a “white Christmas.” I was carefree and gave little thought to the future. It was the second happiest period of my life. Still, since I was good at physics and mathematics, I decided that if I were to attend university, it would be in the Faculty of Science.

2003 – 2009 (Undergraduate and Master’s Student)

As a student, I studied physics and mathematics in the Faculty of Science. At the same time, I began teaching myself programming, learning C, C++, and C#. Around my third year, I became interested in Linux servers and built a home server, creating my first website—a site introducing tourist attractions in Kyoto—by coding HTML and CSS from scratch using Sakura Editor.

During this time, I acquired foundational knowledge of domains, HTTP servers, mail servers, and file servers. I also encountered Linux for the first time and learned SSH, Vim, and Linux commands through hands-on practice.

Although these activities did not earn me academic credit, I vividly remember typing away at my keyboard late into the night. Without these experiences, I would not be working independently today. Even though I have lived only about 38 years so far, I truly believe that one never knows what will prove useful in life.

In parallel, as a physics major, I studied classical mechanics, quantum mechanics, electromagnetism, statistical mechanics, relativity, quantum information, quantum optics, linear algebra, algebra, and analysis. Looking back at my transcripts recently, I realized that my grades in physics courses were quite strong. Attendance was not required at the time; as long as you showed up for exams and passed, you earned credit—so perhaps I simply performed well on tests.

I believe that my ability today to understand machine learning and information science papers down to the mathematical level stems from the effort I put in during this period.

2009 – 2015 (Company Employee – Device Development)

After completing my master’s degree, I joined a manufacturing company and was assigned to a semiconductor device development department. Through development, design, and simulation work, I was able to immerse myself in numerical computation. I deepened my knowledge of the Poisson equation, Boltzmann equation, diffusion equations, finite element methods, and linear algebra, gaining extensive experience in computational physics, including matrix computation optimization and algorithm design. I even worked on CAD and mask design, though I doubt I will ever do so again.

This period also taught me the fundamentals of business, as well as essential professional skills such as Excel usage, PowerPoint presentations, and patent writing. I am deeply grateful to the senior colleagues (who later became my supervisors) who entrusted me with these responsibilities.

2015 – 2019 (Company Employee – System Development)

I later transferred to a systems development department, where my self-taught experience in web application development was recognized. As a project leader, I gained experience developing full-scratch web services deployed worldwide. Until then, my system development experience had been entirely self-taught, so working on large-scale systems with many people was both refreshing and educational.

While large enterprises can be inefficient, in environments where even a single system failure is unacceptable, such structures may be inevitable. This too was a valuable experience.

Through my work, I also learned advanced Vim and terminal techniques from highly skilled colleagues. Although I had already customized my tools to improve efficiency, I realized there was always someone more skilled. The world is vast. Without working alongside such individuals, my current work efficiency would likely be much lower.

2019 – Present (Independent)

I became independent once my system development work at the company had stabilized.

Currently, I work on machine learning model development as well as business system development using GCP and AWS. I have a particular interest in recommender systems. I have also resumed studying quantum information, which I learned as a student, in response to renewed interest in quantum computing.

Developing machine learning models and systems requires not only knowledge of machine learning but also mathematics, computer science, and experience with modern cloud-based systems as a whole.

2021 – Present (Working Ph.D. Program)

In the fall of 2021, I began my doctoral studies as a working professional, focusing on recommender systems. I had already participated in multiple recommender system projects professionally, and as I continued development, I found myself reading more academic papers. This led to a strong desire to study recommender systems more deeply, motivating my decision to pursue a Ph.D.

Summary

Looking back, all the technologies I have worked with since my student days—

  • Physics
  • Mathematics
  • Programming
  • Linux server development
  • Numerical computation
  • System development

—have proven useful in my current work. I did not choose these skills with a long-term plan in mind, but I am glad that I continued to make efforts at each stage of my life. The only experience left unused so far is mahjong—but perhaps that knowledge will become useful in old age.