Christopher Howie

I've spent the last two years figuring out where AI tools actually help an engineering team — and where they quietly get in the way.

The short answer: they help most when the fundamentals are already solid. At Cars Commerce, I integrated Claude Code into production delivery workflows on a platform serving dealer-facing APIs — and the reason it worked wasn't the AI. It was sixteen-plus years of engineering discipline underneath it. On that platform, I shipped React and TypeScript up front and PHP/Laravel behind it, ran Kong API gateway infrastructure on Terraform-provisioned AWS, and owned the Datadog observability stack (dashboards, tracing, alerting) that kept on-call sane.

That's the pattern across my career. I build systems that stay up.

Right now, every engineering org is working out what AI adoption actually looks like in practice. If that's the problem on your desk, it's the one I've been living in — and I'd rather help a team get there than tell them how.