Matt Moats
Fourteen years in the engine room. I started in operations, spent eleven years at Amazon on tier-1 AWS services, and now lead an engineering team at Oracle. I write here about what I am learning — mostly about distributed systems, leadership, and what AI tooling is doing to both.
- → Lead an engineering team on distributed systems, reliability, and developer-platform work.
- → Run hiring, performance, on-call culture, and technical direction. Most days I am writing fewer PRs and more documents.
- → Currently spending an unusually large fraction of my time on AI-assisted development workflows and what they do to team dynamics.
- → Eleven years on tier-1 services. Multi-region, high-throughput, the kind of system where a 9-second flat line gets a serious meeting.
- → Came in through TechOps holding the floor at 3am, moved into systems engineering, then into writing the services I used to operate.
- → Wrote postmortems, ran on-call rotations, and got an unfair education in how production systems actually fail before I ever shipped a feature.
- → Helped launch a AAA MMO. Linux/Windows platform, 24/7 NOC, learned that "live service" is a personality trait, not a job description.
A system that wakes engineers at 3am is a system with a budget problem. Reliability work is feature work in a trench coat.
I push for measurable changes. "Faster" means a graph. "Better" means a SLO. Otherwise you are decorating.
A 30%-slower system that a tired on-call can debug in five minutes wins. Always.
Experience is a starting point. Trajectory is the product. I will take a hungry SDE I over a defended Senior every time.
The kind thing is to tell someone what is not working in week two, not month nine. I learned this the hard way.
TryHackMe on weekends, the occasional side project, and writing. Management makes you stupid in specific ways. Fight it.