Eukarya Inc.

Company

Key Value
Company Eukarya Inc.
Website https://eukarya.io/
Description Eukarya develops an open-source platform. It supports a nationwide 3D digital twin project led by the Japanese Ministry of Land.
Location Tokyo, Japan

Team

Key Value
Title Software Engineer (Backend)
Mission My mission was to develop a workflow engine for an open-source ETL tool.
Term 2024-10 - 2025-08
Type Freelance

Projects

Key Value
Summary EU1. I developed a processor that compiles user scripts into WebAssembly and executes them on a WASM runtime, which enabled users to customize workflows efficiently.
Situation
  • Users wanted to use their own scripts (Python, Go, Rust) to extend the workflow engine.
Task My mission was to make the engine run user scripts safely by compiling them to WebAssembly.
Action
  • I developed the compilation system:
    • I designed a processor that turned user scripts into .wasm files
    • I supported multiple programming languages (Python, Go, Rust)
  • I integrated a WASM runtime to the processor
Result As a result, this feature made the engine more flexible. Users could customize workflows with their preferred languages.
Challenge A key challenge was giving users flexibility while keeping the system secure.
Solution To solve this, I made a clear interface for Wasm modules. It controlled how data passes and how functions are called. This kept user scripts safe.
Learning I learned that I must balance flexibility and security when running user code.
Skill Fast Learning
Key Value
Summary EU2. I developed a real-time logging system with efficient storage and caching, which enabled users to monitor the workflow status instantly from the UI.
Situation
  • Users needed to monitor their workflow status in real-time.
Task My mission was to build a logging system so users could check workflow status instantly from the UI.
Action
  • I developed the logging infrastructure:
    • I made the engine publish logs to Pub/Sub
    • I created a subscriber to fetch logs from Pub/Sub and store them in a datastore
  • I developed the API endpoint:
    • I created a WebSocket endpoint for real-time log streaming.
    • I added a REST API endpoint to fetch historical logs.
Result As a result, users could see worker status in real time. This improved monitoring and troubleshooting.
Challenge A key challenge was designing a logging system that was fast for real-time viewing and cheap for long-term storage.
Solution To solve this:
  • I used Redis cache to make it fast.
  • I used Google Cloud Storage for cheap long-term storage.
Learning I learned that using different storage types - cache for recent data, object storage for old data - balances speed and cost.
Skill Technical Challenge

Technology

Value Tag
Go Backend
GraphQL Backend
MongoDB Backend
Redis Backend
Rust Backend
Google Cloud Infrastructure
Terraform Infrastructure