---
title: "About Ying Xiang Zhao"
url: https://geminixiang.com/about/
description: "DevOps & Backend Engineer with 4+ years across GCP / AWS infrastructure, Kubernetes, Terraform IaC, large-scale data pipelines, ML / LLM model serving and AI agent infrastructure. Currently at GliaCloud running 10 GKE clusters and 200+ serverless services."
---

# About Ying Xiang Zhao

**DevOps Engineer** · Taipei, Taiwan

DevOps & Backend Engineer with 4+ years across GCP / AWS infrastructure, Kubernetes, Terraform IaC, large-scale data pipelines, ML / LLM model serving and AI agent infrastructure. Currently at GliaCloud running 10 GKE clusters and 200+ serverless services.

## Open to work

目前開放新的工作機會（open to work），歡迎來信 contact@geminixiang.com 或透過 Cake / LinkedIn 聯繫。

- Roles: DevOps Engineer, Platform Engineer, Site Reliability Engineer, AI Infrastructure Engineer
- Locations: Taipei, Taiwan, Remote
- Contact: [Email](mailto:contact@geminixiang.com) · [Cake](https://www.cake.me/geminixiang) · [LinkedIn](https://linkedin.com/in/geminixiang/)

## Open source

### mikan

Self-hosted AI coding agent for organizations: every Slack channel and DM gets its own sandboxed workspace, credentials and memory; also speaks Telegram, Discord and GitHub.

[GitHub](https://github.com/geminixiang/mikan) · [Website](https://mikan.geminixiang.com) · @geminixiang/mikan (npm) — TypeScript, Docker, Kubernetes, Helm, Kata Containers

- Built on pi-agent-core with a unified sandbox execution layer for Docker and Kubernetes Agent Sandbox
- Workload isolation through Kata microVMs; conversation-scoped resource and credential boundaries, mount-collision protection, resource limits, idle cleanup and state reconciliation
- Deployed to GKE and k3s with Helm; adopted internally at GliaCloud for ops, debugging automation and data queries

### tiny-agent

Minimal AI coding agents implemented side by side in TypeScript, Go, Python and Rust, with a Traditional Chinese book explaining how coding agents work.

[GitHub](https://github.com/geminixiang/tiny-agent) · [Website](https://tiny-agent.geminixiang.com) — TypeScript, Go, Python, Rust

- Each runtime calls OpenRouter, runs bash / read / write / edit tools, and loads AGENTS.md and skills
- Supports MCP, session resume, context compaction, usage / cache accounting and cancellation
- Default model chosen from a cross-runtime eval of speed, stability, convergence and token efficiency

## Experience

### DevOps & Backend Engineer — 集雅科技 GliaCloud

_2023-02 – present_

**Infrastructure**

- 主導 IaC 實踐，採用 Terraform (Terramate) 維護公司所有雲端基礎建設，並以 AI Agent 加速審核與建置流程，大幅節省人力成本
- 建構大規模用戶資料收集系統，每月處理 3,000 億次請求，涵蓋 Load Balancer、GKE、Cloud Logging、Cloud Storage、BigQuery 完整資料處理 pipeline
- 管理 10 個 GKE cluster、200+ 個 serverless service，確保系統高可用與穩定性

**AI Agent Infrastructure**

- 主導 mikan agent 導入，降低使用門檻，輔助公司運維及管理、系統偵錯自動化、產品進度管理與即時資料查詢

**SaaS Operations**

- 管理 SaaS 專案完整雲端架構：Cloud Run、Cloud SQL、Memorystore for Redis、API Gateway
- 優化對外 API 安全機制，強化身份驗證與存取控管

**Performance**

- 定位 DB slow query 根因，將隨資料量線性增長的 API 回應時間從數分鐘壓縮至 3 秒內
- 針對高併發場景重構 ML 任務批次寫入機制，提升系統吞吐量數倍

**ML Model Deployment**

- 部署並維運多套 LLM 推論服務（自行 fine-tune 語言模型、開源 video / image 生成模型），依服務世代採 PyTorch Serve / NVIDIA Triton / BentoML 三套維運策略，並維護 HuggingFace、Replicate 上的實驗性部署流程
- 在 Azure NVIDIA A10 GPU VM 上建置並維運 NVIDIA Omniverse 基礎設施，支援 3D 場景生成與 AI 視覺內容研究

### Backend & DevOps Engineer — 鈊象電子 IGS

_2022-03 – 2023-02_

- 負責遊戲營運後台 API 開發與系統整合，建構模組化工具框架支援業務資料管理
- 運用 AWS EC2、Amazon MSK、AWS Lambda、Amazon Bedrock 建構雲端服務與資料流處理架構
- 設計並維護 Jenkins CI/CD pipeline，提升部署效率與穩定性
- 與台灣科技大學（NTUST）產學合作，設計遊戲日誌即時資料處理框架

### Frontend Intern — 博錸生技

_2018-08 – 2018-10_

- 執行官網全站 RWD 改版（ASP.NET），提升頁面載入速度與行動裝置體驗

## Other projects

- [SentriPay](https://sentripay.com) — Travel expense splitter I run: multi-currency conversion, no app to install, and a shareable link that keeps everyone's trip ledger in sync.
- [Studio](https://studio.geminixiang.com) — Browser-based video editor built on WebCodecs.

## Skills

- **DevOps / Cloud**: GCP, AWS, Cloudflare, Terraform / CDKTF, Kubernetes, Docker, Cloud Build, GitHub Actions, ELK Stack, SonarQube, Linux, Git
- **Backend**: Python, Express, Bash, MongoDB, PostgreSQL, MySQL, Redis, RabbitMQ, Kafka, Spark
- **ML / AI**: PyTorch Serve, NVIDIA Triton, BentoML, HuggingFace, Replicate, AI agents
- **Frontend**: Vue, React, Astro

## Education

- 國立政治大學 National Chengchi University — Master's, Library & Information Science (2019–2021)
- 國立臺北科技大學 National Taipei University of Technology — Bachelor's, Computer Science (2014–2018)

Languages: Mandarin (native), English

## Profiles

- GitHub: https://github.com/geminixiang
- LinkedIn: https://linkedin.com/in/geminixiang/
- Cake: https://www.cake.me/geminixiang
- Website: https://geminixiang.com

## misc

I'm a DevOps Engineer who loves diving into diverse domains — from full stack web development to ML model deployment, all the way to creating mind-blowing videos with diffusion models. Always chasing that next cool tech to play with.

---

## About the author

**Ying Xiang Zhao** (geminixiang) — DevOps Engineer, Taipei, Taiwan.

Ying Xiang Zhao（geminixiang）是住在台北的 DevOps 工程師，工作橫跨雲原生基礎設施、CI/CD、ML 模型部署與全端開發。這個部落格記錄 infra、資安與 AI coding agent 的實戰筆記；作品包含開源專案 mikan（把 AI coding agent 帶進 Slack / Telegram / Discord）與 tiny-agent（用四種語言實作的極簡 coding agent），瀏覽器內的影片編輯器 Studio，並經營旅遊分帳工具 SentriPay。

**Open to work** — DevOps Engineer / Platform Engineer / Site Reliability Engineer / AI Infrastructure Engineer · Taipei, Taiwan / Remote. Résumé: https://geminixiang.com/about/ (JSON: https://geminixiang.com/resume.json)

Topics: DevOps, Cloud Native, Kubernetes, Docker, Terraform, Google Cloud, CI/CD, Security, AI coding agents, ML model deployment, TypeScript, Python

Projects:
- [mikan](https://github.com/geminixiang/mikan) — Self-hosted AI coding agent for Slack, Telegram, Discord and GitHub with per-conversation sandboxes
- [tiny-agent](https://github.com/geminixiang/tiny-agent) — Minimal AI coding agents in TypeScript, Go, Python and Rust, plus a book on how they work
- [SentriPay](https://sentripay.com) — Travel expense splitter with multi-currency conversion and shareable, live-synced trip ledgers
- [Studio](https://studio.geminixiang.com) — Browser-based video editor built on WebCodecs

[About](https://geminixiang.com/about/) · [Résumé](https://geminixiang.com/about/) · [LinkedIn](https://linkedin.com/in/geminixiang/) · [GitHub](https://github.com/geminixiang) · [Twitter](https://x.com/geminixiang)
