I'm a staff software engineer at eBay. For most of the last decade I've worked on the systems that deliver communications to buyers and sellers at eBay scale: notifications, transactional messaging, the platforms underneath them, and the reliability work that keeps them standing.

These days I spend a lot of my time on two things. The first is production reliability and performance at scale, the kind of problems where capacity models break down, framework assumptions stop holding, and nobody is quite sure why a system built five years ago is suddenly paging three times a week. The second is figuring out how AI agents change on-call work: what happens when Claude has access to your observability stack, your service registry, and a real browser, and can actually investigate a page instead of just talking about it.

Before eBay I worked at Persistent Systems on enterprise data platforms and storage systems. I've spent time on Hadoop clusters for large US retailers, on performance tuning for storage manufacturers, and on the kind of unglamorous infrastructure work that quietly moves terabytes while everyone argues about the frontend.

This site is where I write things down. The goal is to share what I learn about building and running large systems without burying the hard parts. If something here is useful to you, or wrong, I'd like to hear about it.

Reach me on LinkedIn or X (@apsunde).