Available for opportunities

Yuan
Genggeng

Data Engineer / Backend Engineer

Databricks & Lakehouse · 6+ yrs distributed systems · Tokyo

legacy_modernization.diff+2 −2

@@ insurance core — NTT DATA @@

- Struts 1.x · JDK 1.7 · Oracle 11g

+ Spring MVC · JDK 17 · Oracle 18c

@@ membership DB — Rakuten @@

- Couchbase

+ Cassandra · 300M records

Who I Am

Senior Backend & Data Engineer with 6+ years of specialized experience in large-scale distributed systems, currently actively learning the modern Databricks Lakehouse ecosystem.

Continuously expanding end-to-end data engineering capabilities, backed by a solid foundation in the Spring Ecosystem, high-concurrency database migrations, and 20 years of software engineering.

Proficient in modern full-stack development (Next.js / Convex / FastAPI) and highly efficient in AI-augmented workflows using Claude Code, Codex, and Antigravity.

🇨🇳 ChineseNative
🇯🇵 JapaneseBusiness
🇺🇸 EnglishBusiness
0
+
Years in Software
0
+
Years Backend Specialist
0
M
Records Migrated
0
+
Languages & Frameworks

What I've Built

Olist Lakehouse Pipeline

Production-Grade Incremental Medallion Lakehouse

An end-to-end incremental Lakehouse on Databricks processing 100k+ real Brazilian e-commerce orders, featuring 3 Kimball fact types, CDF-driven Gold, and Lakeflow DLT pipelines.

  • 🧱 3 Kimball Fact Table Types (Transaction, Accumulating & Periodic Snapshots)
  • ⚡ Delta Lake Change Data Feed (CDF) for ~1% incremental Gold processing
  • 🔄 Dual-Paradigm: Imperative PySpark vs. Declarative Lakeflow (DLT)
  • 🛡️ Automated DQ Gate (<5% threshold) & 64 automated E2E/Unit tests
DatabricksPySparkDelta LakeAuto LoaderSCD Type 2Asset Bundles

PeraPera

High-Performance Japanese Learning Ecosystem

A hybrid multi-service platform with Meilisearch powering search across millions of records. Features a multi-tier AI fallback architecture and Stripe subscriptions.

  • 🔍 Meilisearch for millions of records
  • 🛡 Resilient multi-tier LLM fallback
  • 💳 Stripe subscription integration
  • 🚀 CI/CD to Azure & Railway
Next.jsConvexFastAPIMeilisearchStripeAzure

PeraTube

Japanese Video Example-Sentence Search API

A developer-facing SaaS that indexes massive YouTube subtitle corpora into searchable segments — query how Japanese is actually used in real videos, with millisecond latency and quota-based billing.

  • 🔎 Millisecond search over subtitle segments
  • 🔑 API keys · atomic quota billing via Convex
  • 💳 Stripe subscriptions + credit packs
  • 🧪 15+ pytest suites · Railway + Cloudflare
FastAPINext.jsConvexMeilisearchStripeRailway

Jessie Signal

Real-Time Market Sentiment Dashboard

A React/Convex serverless web app aggregating live market data (VIX, CNN Fear & Greed, major ETFs) and classifying the market into actionable sentiment scenarios.

  • 📊 VIX + Fear & Greed live data
  • 🔔 Email alerts on scenario changes
  • ⚡ Server-side caching for instant loads
  • 🔐 Gated admin dashboard (Clerk auth)
ReactConvexClerkResendRailway

What I Work With

Data Engineering & Lakehouse

DatabricksPySparkDelta LakeLakeflow Pipelines (DLT)Unity CatalogMedallion ArchitectureSCD Type 2CDC & Streaming

Spring Ecosystem

Spring Boot (2.x/3.x)Spring MVCSpring SecuritySpring Data JPA/MyBatisSpring Core (DI/IoC)

Backend Development

JavaKotlinPython (FastAPI)GoMicroservicesDDDRESTful APIsVert.x

Search & Data

MeilisearchKafka (CDC)CassandraCouchbaseSnowflakeOracleRedisRabbitMQ

Cloud & DevOps

Azure (AKS)RailwayAWSDockerPodmanMinikubeGitHub ActionsGitLab CI/CD

AI Tools

Claude CodeAntigravityCodexDeepSeekAI-Augmented Workflows

Frontend & Full-Stack

Next.jsReact.jsVue.jsConvexTailwind CSSTypeScript

Professional Experience

NTT DATA

Contract Backend Engineer

Apr 2025 — Aug 2026📍 Tokyo, Japan
Insurance System Modernization (Struts → Spring)
  • Modernizing a core insurance application from TERASOLUNA 2.2.0 (Struts 1.x) / JDK 1.7 to TERASOLUNA 5.x (Spring MVC) / JDK 17
  • Phase 1: Refactored legacy JSP templates and intra-mart custom tags to improve front-end usability
  • Phase 2: Refactored business logic into clean, decoupled, testable Spring-managed beans using Dependency Injection
Java 17Spring MVCTERASOLUNA 5.xintra-martJSPOracle

Rakuten Group, Inc.

Contract Backend Engineer

Sep 2021 — Mar 2025📍 Tokyo, Japan
Membership System & Database Migration
  • Ran a high-concurrency PoC on Couchbase, generating 300 million mock membership records for benchmarking
  • Executed the migration from Couchbase to Cassandra after stress tests revealed degradation during node rebalancing
  • Built an async Vert.x benchmarking API server and ran distributed Gatling stress tests via GitLab CI/CD
  • Packaged a custom Argon2 hashing library and automated API test suites with Postman/Newman on MiniKube
KotlinVert.xCassandraCouchbaseGatlingGitLab CI/CD
Rakuten Fashion — Agency Order Management
  • Built Spring Boot @Scheduled batch tasks with RabbitMQ to route stalled orders back for asynchronous re-processing
  • Streamlined Gradle builds and API version management across microservices with JFrog Artifactory
  • Standardized API specs with OpenAPI (Swagger) across all backend services
Spring BootRabbitMQGradleJFrog ArtifactorySwagger

Earlier Roles

Software Engineer & Game Client Programmer

2002 — 2021
  • Techoes, Tokyo (2020–2021): Multiplayer game backend — online chat & AI pathfinding with Go, gRPC, Protobuf; live streaming platform (Vue.js, PHP/Laravel)
  • NEC, Tokyo (2018–2019): VMS integrations for the National Police Agency (C# / Milestone XProtect); Mobile Suica backend APIs & batch tasks
  • Game Dev, Shanghai (2003–2018): Client-side development in C++/C#/Unity/Unreal across studios incl. 2K Games China — combat systems, UI frameworks, performance optimization

Let's Work Together

Currently based in Tokyo, Japan. Open to backend engineering opportunities, consulting, and interesting side projects.

Email
kidyuan@foxmail.com
Phone
070-4017-5377
Location
Tokyo, Japan
Visa: Engineer / Specialist (3yr)