About
I’m cloudplz — principal cloud infrastructure engineer. Eighteen years across cloud, infrastructure as code, and database platforms. The domain is data infrastructure; the discipline is platform engineering. Everything here is written from production, not from documentation.
The arc
I started in database operations: MySQL backups, replication, capacity planning, 3 AM pages. Then Vitess, when horizontal scaling in MySQL meant running a distributed system yourself — three years of production operations that taught me what the sharding tax actually costs.
When the org moved to AWS, I led database migrations: about fifty production databases to Aurora over two years, most cutovers under three minutes of downtime, including multi-terabyte tier-0 systems with 24/7 write traffic. Along the way: Aurora Global Database with cross-region replication lag cut from seconds to ~70ms, and SOC2 audits that came back with no findings.
Then the shift from operating infrastructure to building products out of it. That’s where the interesting problems turned out to live.
What I’ve built
- No-code infrastructure platform — product teams provision production-ready Aurora clusters through Backstage templates without writing Terraform. Roughly 80–95% faster than the ticket queue it replaced, an estimated 3,000+ engineering hours a year given back.
- Terraform module ecosystem — curated modules with 8,000+ usages across 70+ repositories. The compliant, production-hardened way is also the easiest way.
- Database canary applications — customer-zero validation across seven data technologies, so platform changes break in the canary, not in a customer’s incident channel.
- AI-enhanced infrastructure tooling — RAG, MCP, and LLM-agent workflows applied to infrastructure operations. Agents are infrastructure; I provision them like it.
The through-line: infrastructure that works without me. When teams can provision what they need without asking, the platform team finally has time to improve the platform.
What I write about
The platform layer that hundreds of engineers build on: Aurora and Vitess scaling decisions, Terraform module design, self-service infrastructure, and what breaks when these meet production. Written for people building platforms, not spinning up their first database.
If a post would make sense on the AWS blog without the platform perspective, it doesn’t belong here.