I'm a software engineer based in Japan, primarily focused on backend engineering.

I enjoy understanding not only how to make software work, but also why systems are designed the way they are:

• How should responsibilities be separated?
• How should data move through an application?
• How should failures be handled?
• How can a system remain testable and maintainable as it grows?
• How does a technical decision affect the user experience?

I prefer learning technologies by integrating them into real applications and improving those applications through small, reviewable changes.

Evolving a simple ASP.NET MVC CRUD application into a structured and maintainable backend system.

This project started as a small CRUD application and is being incrementally improved to understand why common backend architecture and infrastructure components are introduced.

Architecture · Persistence · Dependency Injection
Testing · Error Handling · Maintainability

• Controller → Service → Repository architecture
• Repository abstraction
• Unity Dependency Injection
• Entity Framework 6 and SQL Server persistence
• Code First Migrations
• DTO and Entity separation
• AutoMapper
• Service-layer validation
• Global HTTP 400 / 404 / 500 exception handling
• Application logging with log4net
• MSTest service-layer unit tests
• Vue.js integration

Implemented server-side pagination with:

• LINQ OrderByDescending, Skip, and Take
• Pagination request and response DTOs
• Total count and pagination metadata
• Invalid parameter and out-of-range handling
• Previous/Next controls in Vue.js
• Boundary and validation tests

Development trail:

Issue #26
→ Merged PR #27

C# · ASP.NET MVC 5 · .NET Framework 4.8
Entity Framework 6 · SQL Server · Vue.js · MSTest

➡️ Explore Dotnet Backend Study

Connecting people through shared meals with a multi-service social dining platform.

Let Eat Go is a four-person team project organized across three repositories: a Next.js web client, NestJS backend API, and FastAPI AI service.

• Social dining discovery, hosting, and participation
• Google and Kakao OAuth authentication
• Kakao Maps-based event exploration
• Socket.IO real-time chat
• Albums, comments, likes, and reviews
• DistilBERT-based inappropriate-text classification
• Docker and GitHub Actions-based AWS deployment

Next.js · TypeScript · NestJS · PostgreSQL · PostGIS Socket.IO · FastAPI · DistilBERT · Docker · AWS

Because this was a team project, the organization page documents the architecture, repository responsibilities, contributors, and current portfolio reconstruction status.

➡️ Explore Let Eat Go

Helping seniors and their families stay connected through asynchronous video communication.

Tsunagaroom is a Java/JSP team project designed around a simple problem:

Traditional video calls require both sides to be available at the same time.

The application instead delivers family videos to senior users and automatically records their reactions during playback.

Family records a video
        ↓
Senior plays an unread video
        ↓
Camera and microphone recording starts
        ↓
Playback ends
        ↓
Reaction video is uploaded
        ↓
Original video is marked as read

• Designed the interaction between playback and automatic reaction recording
• Structured the Servlet → Logic → DAO processing flow
• Designed unread/read video-state management
• Implemented and reviewed parts of the recording/upload workflow
• Participated in screen, server, and database design
• Coordinated specifications and implementation decisions within the team

Because this was a team project, the repository clearly distinguishes my contributions from the work of the full team.

Java 21 · Jakarta Servlet · JSP
JavaScript · MySQL 8 · Apache Tomcat 10

➡️ Explore Tsunagaroom

Reconstructing a 2024 AI hackathon prototype as a privacy-safe and reproducible software project.

The original prototype was developed during an intensive AI program in Ulm, Germany and received a hackathon award as a team project.

The current repository is an independent reconstruction using synthetic CRM data. It does not contain the original corporate dataset, proprietary code, internal documents, or credentials.

Synthetic CRM Data
        ↓
Feature Pipeline
        ↓
ML Prediction API
        ↓
Explanation & Evidence
        ↓
React Dashboard
        ↓
Human-reviewed LLM Draft

• Leakage-safe machine-learning evaluation
• Reproducible data and feature pipelines
• FastAPI prediction and explanation endpoints
• React and TypeScript dashboard
• Human-reviewed LLM-assisted outreach
• Automated tests, Docker, and GitHub Actions

🚧 Current phase: repository and safety foundation, including a minimal API health endpoint and automated test.

➡️ Follow the Reconstruction

Technologies used across personal projects, team development, professional practice, and international training.

C# · .NET · Java · TypeScript · JavaScript · SQL

Primary

ASP.NET MVC 5 · .NET Framework 4.8 · Entity Framework 6
JSP · Jakarta Servlet

Additional Project Experience

Node.js · NestJS · FastAPI · Ruby on Rails

Vue.js · React · Next.js · TypeScript
Vite · Tailwind CSS · Razor · HTML · CSS

SQL Server · PostgreSQL · PostGIS
MySQL · SQLite · TiDB · Supabase

Additional experience:

Redis · Meilisearch · Entity Framework · Active Record

Docker · AWS · Azure · Linux
MSTest · Selenium · GitHub Actions

Development practices:

GitHub Issues · Feature Branches · Pull Requests
Code Review · Incremental Refactoring

Machine Learning

K-Means · KNN · Logistic Regression
Decision Tree · Random Forest

Deep Learning & Generative AI

PyTorch · CNN · RNN · Transformer
Azure OpenAI · LangChain

flowchart LR
    A["Build Features"]
    --> B["Design Boundaries"]
    --> C["Improve Reliability"]
    --> D["Scale with Cloud"]
    --> E["Apply AI"]

    style C fill:#d1fae5

My current focus is moving from simply implementing features toward understanding the larger engineering concerns around them.

Feature Implementation
        ↓
Application Architecture
        ↓
Database & Persistence
        ↓
Testing & Reliability
        ↓
Cloud Infrastructure
        ↓
AI-enabled Systems

Learning, working, and adapting across different countries and cultures.

These experiences have strengthened my adaptability, cross-cultural communication, and ability to work in unfamiliar environments.

I learn best by improving real applications through small, reviewable changes. When introducing a technology or pattern, I focus on why it is needed, what problem it solves, where it belongs, and what trade-offs it introduces.

My goal is to understand both implementation details and the larger systems around them.