The Senior Engineer Who Googles Everything

I’m Leon Shelhamer, and I’ve been building software professionally for sixteen years. It took me most of those years to be able to say this next part out loud without wincing: I don’t have the vocabulary memorized.

Ask me to stand at a whiteboard and define a monad, or recite the four pillars of object-oriented programming in the right order, or explain from memory the precise difference between a mutex and a semaphore, and I’ll stall. I’ll know the shape of the answer. I’ll know when I’m looking at one in the wild. But the textbook word for it isn’t sitting in the front of my head, ready to recite.

For a long time I thought that disqualified me. I came into this industry through the side door - a high school dropout with a GED who spent eleven years working restaurants and bars around Albany before I ever opened a code editor. Nobody handed me the curriculum. I taught myself one Google search at a time, reverse-engineering WordPress core line by line until the machine stopped being a mystery. I learned what things did long before I learned what they were called.

What the job actually rewards

Here’s what sixteen years taught me: the senior part of this work was never the vocabulary. It’s the judgment. It’s knowing which decisions are load-bearing and which ones don’t matter. It’s looking at a system and feeling, almost physically, where it’s going to crack under load before anyone’s written a line of the thing.

I can’t always tell you the academic name for the pattern I’m reaching for. But I can tell you it’s the right pattern, why the obvious alternative will hurt you in eight months, and what it’ll cost to change your mind later. That’s the part you can’t Google in the moment - and it turns out it’s the part that actually separates a senior engineer from a junior one.

AI didn’t replace what I know - it erased what I forget

For most of my career, my weakness was the recall gap. I’d know exactly what I wanted to build and lose ten minutes hunting for the exact function signature, the exact flag, the exact name of the thing. The architecture was instant; the trivia was the tax.

I’ve written before about the other half of this - being a lifelong slow typist, and how voice tools finally got my hands out of the way. That was about getting the words out. This is the opposite bottleneck: not having the word in to begin with. Same career-long friction, different end of the pipe.

AI closed that gap to nearly zero. Now the recall is free and instant, which means the only thing left on the table is the judgment - and judgment is the thing I actually have. I describe the system I see in my head, the assistant supplies the syntax and the terminology, and I spend my attention on the decisions that matter instead of on remembering what the decisions are called.

People worry that AI makes engineers interchangeable. My experience is the opposite. It strips away the part of the job that rewarded memorization and leaves the part that rewards taste. If you’re the kind of engineer whose value was “I memorized the docs,” that’s a problem. If your value was always “I can see how the whole thing fits together,” this is the best era you’ll ever work in.

The point

I will probably Google a syntax I’ve used a thousand times, on the same day I quietly talk a team out of an architecture that would have cost them a quarter. Both of those are true. Only one of them matters.

I’m not the engineer who knows every word. I’m the engineer who knows where the bodies are buried - and these days I have a tool that hands me the words while I do the part that’s actually hard.

Leon Shelhamer is a senior software engineer with sixteen years of experience and the developer behind Two Eleven Jay and the managed-hosting service uphost.ly. He works in React and Vue, runs his own production infrastructure, and builds AI-augmented development workflows from upstate New York. Connect on LinkedIn.