From Vibe Coding to Vibe Engineering
AI systems from an engineering org’s point of view
Welcome to Engineered Intelligence — a newsletter about building AI systems like an engineer, not just a model-wrangler(sic wrapper). This is where software meets AI from an Engineering Organization’s perspective: architecture, deployment, team structure and everything in between.
A while back, I wrote about vibe coding. It's fast, flowy, experimental mode of building where you don't overthink things. You just try it. Decide from the browser, console or wherever you can see the results. Don’t over analyze the middle. Glue together a POC with whatever tools the model has lying around. It’s intoxicating, productive, and often gets you to something cool, fast.
But here’s the thing: vibe coding doesn’t scale with large projects.
It breaks under the weight of actual users, teams, actual expectations, and the reality of production systems.
So what happens next?
That’s where vibe engineering comes in.
Vibe engineering is the slower, more deliberate mindset shift that happens when the stakes go up. You start asking different questions. You’re no longer building for yourself. You’re building with a team, for reliability, for maintainability and for longevity.
This sentiment more than anything else feels like what working with AI on production systems and in Engineering teams is like today. We’ve sped up the act of writing software. That’s no longer the bottleneck, we’ve just accelerated to the next bottlenecks.
This newsletter is for engineers and leaders navigating the messy middle of AI integration.
You'll get:
Deep dives into how we're using AI on a larger team
Implementation notes, good and bad
What we're learning from working with our CX clients at XtendOps
Playbooks for AI system design, from prompt to production
Tactics for getting non-technical stakeholders on board
Think of this as your dispatch from the real world of AI Engineering—no hype, just builds and build failures.
A bit about me: I lead a team of developers at XtendOps. We've built AI tools that interact with real customers, sales teams, support agents, and a whole lot of unknowns. We’ve made mistakes. We’ve shipped wins. This is where I share both.
So hit subscribe to follow along. And if you’re already doing this work—I want to hear from you. Tell me what you’re struggling with. What you’ve figured out.
If this resonated, forward it to a teammate. Let’s get better at building AI together.