Resume

Professional Summary

Full-stack engineer (Rails + TypeScript/Node) with ~10 years building and operating cloud services and data-integration systems. Strong in production reliability: monitoring, incident debugging, and making fragile external integrations stable. Most recently shipped and maintained a family-trust accounting product end-to-end at Trinity-Technology (Figma → Rails backend → web frontend → iOS → production).

Skills

Technical Skills

Experience

Trinity-Technology — Full-Stack Developer (Web/Backend + iOS)

Aug 2021 – Present | Tokyo, Japan

Team: 5 devs, 1 designer, 1 PM (Kanban-ish)  |  Tech: Ruby/Rails, AWS, Heroku, PostgreSQL, GoodJob, Sentry, Scout APM

Built customer-facing accounting software for family trusts to track asset status and transactions and reduce misuse risk.

  • Shipped “Oyatoko” end-to-end: Rails backend, web UI, and iOS app for customer asset visibility and management.
  • Improved production stability using Sentry + Scout APM to identify recurring errors and performance bottlenecks and drive fixes.
  • Built internal log-query tooling to speed incident triage and diagnosis.
  • Ran zero-downtime schema/data migrations across releases with staged playbooks (backfill → deploy → cleanup).
  • Led v2 rebuild of an internal sales onboarding system; redesigned data entry for assets/family structure to improve searchability and speed.
  • Automated Google Slides + legal document generation from customer asset data with conditional clauses and external-lawyer access controls; cut manual prep by ~10 hours/week.
  • Mentored a junior engineer on Git/PR workflow, environment setup, and day-to-day delivery.

Moneytree — Software Engineer

Jan 2020 – Apr 2021 | Tokyo, Japan

Tech: AWS (Step Functions, S3), TypeScript, Node.js, Puppeteer, Ruby, legacy XUL/JS, Sumo Logic

Supported a financial data aggregation platform running ~500k jobs/day on average. Automated browser sessions captured raw HTML/JSON to S3, extracted/normalized data, and sent results to a Rails backend. Debugging relied on logging and sandbox repro (screenshots/data dumps) due to frequent upstream UI changes.

  • Monitored failure rates by institution via log analytics, investigated spikes, shipped fixes, and reran failed jobs to restore ingestion.
  • Built a hackathon incident-debug tool that replayed failed runs from stored artifacts (steps, selectors/actions, network requests, screenshots) and compared against last-known-good runs.
  • Added XHR-based extraction to the Puppeteer framework (instead of DOM selectors) to reduce breakage for SPA targets.
  • Implemented/extended priority-aware retry + backoff to handle rate limits (premium tier + time-since-refresh), reducing queue pileups and cascading failures.
  • Worked with Customer Support, backend API teams, and mobile teams to gather repro details and coordinate cross-system changes.

FUNDBOOK — Web Developer

Apr 2017 – Jan 2020 | Tokyo, Japan

Tech: Ruby/Rails, React, TypeScript, PostgreSQL, RSpec, Cypress, Selenium

Joined early and helped build the product from the ground up: customer web app for M&A listings plus internal tools for the sales team.

  • Built the customer React SPA (search/filter listings, save, inquiries).
  • Built internal tools for tracking buyer activity and managing outreach and responses.
  • Implemented Rails APIs, background jobs, reporting/search; maintained RSpec + Cypress coverage.
  • Built a Selenium-based cross-browser/cross-viewport acceptance testing setup to reduce manual QA and regressions.

Jan 2015 – Jan 2017 | Tokyo, Japan

Team: 4 developers  |  Tech: Ruby/Rails, AWS (S3-triggered workflows), PostgreSQL, MongoDB, mitmproxy (plus MySQL/Redis/Elasticsearch)

Built a system to manage listings, guests, and messaging across Airbnb, Booking.com, and HomeAway. Supported ~300 listings with hundreds of bookings/day and messages/day, replacing manual ops.

  • Built a shared integration layer for multiple booking platforms using a common domain model and platform-specific adapters.
  • Built crawlers/HTTP clients for platforms without public APIs using authenticated web sessions and traffic inspection tools (e.g., mitmproxy).
  • Implemented event-driven automation: customer emails → AWS → S3 storage → parsing → server notifications to trigger actions (message received, booking confirmation).
  • Kept web automation reliable by detecting failure signals and responding quickly to upstream UI changes.
  • Stored raw scraped data in MongoDB and persisted normalized entities in PostgreSQL to preserve traceability.
  • Replaced a manual workflow (email client + per-platform logins) with unified tooling to cut operational overhead.

Additional