English · 中文
SRE at Alibaba Cloud, based in Shenzhen. On GitHub since 2010.
By day I keep cloud-native infrastructure alive. On the side I build tools and knowledge bases — most of them written to be read twice: once by people, once by agents as RAG corpora.
- agent-up — a feedback → attribution → release → rollback loop for improving agents in production
- KUDIG / ResolveAgent — turning runtime traces of ops tools into structured, agent-readable knowledge bases
- open-geo — notes and research on Generative Engine Optimization (GEO)
Applications
| Project | What it does |
|---|---|
| FDE Scope | A complete FDE SOP: 4 zones, 18 stages, 10 gates |
| agent-up | Continuous-improvement loop for agents: feedback, attribution, release, rollback |
| ResolveAgent | Alert in, four tracks diverge, terminal station is root cause |
| Klaw | One binary that holds an entire Kubernetes ops practice |
Open tools
| Project | What it does |
|---|---|
| OpenArch | Presales-architect work as explicit skills and workflows |
| open-tam | TAM troubleshooting distilled into three-part root-cause reports |
| OpenDemo | 959+ runnable, production-grade demos for solution architects |
| oh-my-browser | Everything about your bookmarks |
Knowledge bases
| Project | What it does |
|---|---|
| KUDIG Database | K8s + AI Infra ops knowledge base, written for humans and agents |
| AI Guru Database | GTM and launch playbooks |
| podcast-database | Podcasts for LeetCode refugees — listen until it clicks |
Full index of 100+ repos across 20+ orgs: allengaller.github.io/repos
- allengaller.github.io — home base
- Projects · Repos · Links
Good fits: Agent / RAG / MCP / CLI tools that need to survive production; hands-on design around Kubernetes, observability, and failure analysis; turning scattered docs into knowledge bases that keep growing. Fastest way to reach me is a GitHub Issue on the relevant repo.
© Allen Galler · Shenzhen