The New Bar for Engineers in 2025: AI-Native or Behind

I’m doing a lot of hiring over the next few months, and as I prep for it, I’ve found myself stepping back to ask: What do I actually expect from great engineers today?

Even six months ago, I was just starting to ask candidates about AI tools — whether they’d tried ChatGPT, Cursor, Claude, Copilot, whatever. That already feels dated. The bar has moved. What “good” looks like has shifted again. And I think it’s going to keep shifting fast.

What Hasn’t Changed

Before I get into what’s new, I’ll call out a few things that still matter to me.

I work at a product company, and I still care deeply about hiring product-minded engineers. I want people who understand the why, not just the what. Engineers who think in systems, who know how to navigate trade-offs, and who have experience building and maintaining real products.

Depending on your level (I generally am hiring Senior or higher level roles), I expect some fluency in system design, an ability to write clean code, and a working understanding of when theory matters and when it doesn’t. But more than anything, I care about actual experience — have you launched things? Fixed real problems in production? Owned outcomes?

All of that is still foundational. But it’s not the differentiator anymore.

What’s Changed: AI-Native Is the New Bar

The thing I’m paying the most attention to now is whether you’re an AI-native engineer.

I don’t mean someone who’s just curious about AI or played around with it once. I mean someone who’s actually using these tools — Cursor, Windsurf, Claude Code, Copilot, you name it — in their day-to-day workflow.

If you’re not, I think you’re at a disadvantage. Engineers who use these tools effectively are writing more code, faster, with fewer errors — and often thinking more clearly, too.

That said, just using AI isn’t enough. What matters is how you use it.

Can you explain the code it gives you? Can you modify it? Push back when it’s wrong? Talk me through the decision-making behind it? When I do interviews, I don’t care if you’re typing every character yourself. I care if you can explain what the code is doing, why it works, and where it fits in.

That’s the real skill: pairing technical judgment with leverage.

It’s Not Just About Code

I also care a lot about how engineers are using AI in areas beyond just writing functions.

Do you use ChatGPT or Claude to explore ideas when you’re planning a system? Do you use it to generate diagrams or draft technical docs? Are you turning voice notes or rough outlines into clean, well-structured documents?

Everyone has access to these tools now. But not everyone uses them well. In my experience, what separates great output from average output is judgment — knowing what “good” looks like before you even start. If you’ve written strong PRDs, tech specs, or launch plans in the past, AI will make you faster. If you haven’t, you’re more likely to just accept whatever it spits out.

That’s the distinction I’m trying to assess.

What I’m Looking For in 2025

A few things I’m especially drawn to right now:

• Zero-to-one experience. Have you built something from scratch, in a messy or ambiguous environment, where the stakes were high? That’s hard to fake.
• Product instinct. Do you care about what you’re building? Not just the backend mechanics, but the user experience, the polish, the value?
• Good judgment with tools. Are you someone who can use AI and other modern tooling to increase output, while still understanding what quality looks like?
• Clear thinking, clearly expressed. Whether it’s a doc, a diagram, or a Slack message — can you communicate complex ideas simply?

How I Interview Engineers

I believe the best interviews are designed to get signal from real experience — not from memorized trivia.

I use a coding exercise where you’re welcome to use AI tools. I actually hope you do. What I care about is whether you understand what’s happening, and whether you can talk me through your approach.

I use behavioral interviews to dig into past work: what you built, how you led, what you learned, what went sideways. That’s where the real substance comes out.

And I include live, conversational system design interviews that are meant to feel more like working sessions. These are less about facts and more about your thinking — your opinions, trade-offs, and reasoning. I’ve found those conversations give the clearest view into how someone approaches their craft.

Final Thought

When I’m hiring, I’m not looking for perfection. I am looking for people who are evolving with the field — people who bring experience, judgment, and curiosity, and who use every tool available to build better software, faster.