I’ve spent the last several years building AI systems at Cisco and managing the teams that build them. Most of what I’ve learned about doing that well lives in meeting notes, postmortems, and the heads of people I’ve worked with.
That’s a problem. Knowledge that can’t be written down is knowledge that doesn’t compound.
So this blog is partly writing practice — forcing myself to articulate things I currently only half-understand — and partly a public record of decisions, tradeoffs, and lessons from working at the intersection of engineering leadership and production AI systems.
I’m not writing for search traffic. I’m writing for the engineering manager who is about to make the same expensive mistake I already made, or the team lead trying to figure out how to talk to executives about why their agentic AI pilot keeps breaking.
If that’s you, welcome. I’ll try to be specific enough to be useful.