CI Gates, Test Result Caches, and the Code Graph That Might Tell Us What to Test
What I learned from splitting Rails CI by risk, caching exact test results, and exploring whether AI and code graphs can make test selection safer.
Joey Wang · Software engineer and consultant
I help teams ship and operate software with confidence.
DevOps and reliability, AI-native engineering, and full stack delivery, grounded in evidence from more than two decades of running real systems.
Where I help
DevOps and Reliability
Systems that stay up, deploy safely, and can be explained when they fail.
- Zombie PostgreSQL replicas: the WAL segment removed deadlock
- I Audited the OpenVPN Server I Wrote a Guide For
- Kaniko: building container images without a Docker daemon
AI-Native Engineering
AI that supports engineering judgment, with verification, privacy, and approval built in.
- LLM Strengths and Limitations: A Practical Framework
- AI Application Architecture: LLM + Memory + Tools
- How an LLM Coding Agent Actually Builds Software
Full Stack Development
Long-lived web applications improved incrementally, without pretending a rewrite is free.
- When a Rails or LMS Platform Becomes Hard to Change
- Ruby Ractor vs the GVL: When You Get Real Parallelism
- Rails in Kubernetes: Migrations, Jobs, and Puma
26+
years building and operating web software
254
published articles with reasoning and verification shown
Selected open source
Latest writing
All writingWhen a Green Deployment Still Serves a 500
A REX production incident exposed the gap between a successful Kubernetes rollout and a working application, and what we changed to close it.
The One-Person Company Is a Coordination Problem
I am exploring how a one-person company can use agents, observability, and market intelligence without turning automation into an ungoverned second company.
When Productivity Turns Against Its Purpose
The AI boom promises abundance, but without a fairer distribution of power it may deepen the poverty, dependency, and cultural depletion it claims to overcome.
When the Report Button Becomes a Competitive Weapon
Unfair competition has learned to exploit platforms, moderation systems, and the gap between takedown and appeal, from invented users to AI-generated consensus.
Hermes vs OpenClaw, NanoClaw, ZeroClaw, IronClaw: Agent Architecture
A practical, architectural comparison of Hermes, OpenClaw, NanoClaw, ZeroClaw, and IronClaw, and why a layered agent system beats picking one winner.
Have a platform that has become hard to change?
A bounded audit produces findings, priorities, and a plan before anyone commits to a rewrite.
See the consulting offer