About

I am Erdinc Kaya. I work at HPE in London, on infrastructure for AI workloads — Kubernetes, data platforms, and the operational details that decide whether a design survives contact with a real cluster.

This site is where I write up things I have built and what running them taught me. Most of it starts as a demo I can point at, rather than an argument I want to make. The code for nearly everything here is on GitHub.

Recurring subjects:

• Running AI agents under isolation the cluster enforces, rather than the prompt
• Data platforms, medallion architectures, and moving data between core and edge
• The gap between a control being installed and that control actually working

Views here are my own, not my employer’s.

Elsewhere

• GitHub