6 public issues.
My LinkedIn profile still says "tech junkie that thrives under pressure." Underneath that line is a long attempt to unload responsibility correctly: scripts, abandoned interfaces, GitOps, an internal testbed, and finally a platform I could teach to the development team.
Career archaeology Read issueThe useful signal was not which agent looked best on visible tests. It was which one kept Archibus's boring rules intact after the hidden checks showed up.
Agent evals Read issueThe container is the easy part. Real Archibus workspaces also need databases, licenses, restore paths, startup scripts, IDE access, diagnostics, and support evidence.
Managed workspaces Read issueArchiBot is a repeatable Archibus lab with AI attached. That means the token math has to be honest from the start.
ArchiBot economics Read issueISM is not a generic software shop. Most of the work lives in the hard middle around Archibus, operations, data, support, and tools that make the work repeatable.
ISM Services Read issueThe first useful ArchiBot knowledge-system lesson was not a vector database. It was learning how to make private software knowledge retrievable at all.
RAG archaeology Read issue