Modern engineering leadership
I’ve been thinking a lot lately about what it means to be an engineering leader right now. Some of the things we anchored our careers to are shifting under our feet, and I keep turning over what that means for the people I lead.
There’s a lot of noise about tools writing code for us. Copilots and agents and whatever comes out by the time this gets published. I don’t think any of it changes what teams need from engineering leaders, though it does make the gaps a lot easier to see.
If I’m being honest, coding was never the hardest part of software development. It was just the most visible one. The hard parts were always human. You had to decide what not to build and explain why. You had to help people grow without crushing their confidence, and ship something imperfect and stand behind it. Code was how we expressed all of that, and AI makes those parts a lot harder to ignore.
When craft becomes identity
One of my engineers told me recently that he was hesitant to lean too heavily on AI-assisted tooling. Not because it didn’t work, but because it worked really well. He was worried that the instincts he’d built over years of debugging and shipping would dull, and that he’d lose the thing he built his career on.
He’s not someone who’s afraid of learning new things, and I don’t think his fear is irrational at all.
For a lot of engineers, our value was tied directly to what we could produce with our own hands and brains. We learned through repetition and failure and earned fluency the hard way. When a tool shows up that can do parts of that work instantly, you start asking yourself what you’re actually good at.
That conversation stuck with me because it was really about trust, and whether ten years of experience still counts when a machine can make the same suggestion in a second.
Coding was never the bottleneck
The code was never the bottleneck. It just felt like one because it was slow and manual. What actually slowed teams down was bad decisions and unclear ownership. AI doesn’t fix any of that, and sometimes it makes it worse because bad decisions get built faster.
I was taught that the idea of a “10x engineer” at the leadership level was about creating an environment of mentorship and autonomy. I still believe in hiring smart people and getting out of their way.
What I didn’t expect is how much time I now spend setting shared approaches to tooling and workflows. Without some guardrails, reviews get harder and knowledge stops moving between people. It took me a while to accept, but my team now holds one code review bar for everything that ships, whether a person or a tool wrote it.
The unglamorous part
It doesn’t sound glamorous to say that part of my job is making sure we’re using tools consistently. It sounds like process. But in an AI-assisted world, alignment is where the leverage lives.
When everyone uses tools differently, results vary wildly and quality turns into a matter of opinion. Consistency means people can spend their energy on the actual problem.
Tooling decisions end up being a kind of mentorship. They tell the team what we value, and they help newer engineers learn how to think about problems. If you’re an engineering leader who feels a little lost right now, this is where I’d start.
Decisions got more expensive
When code is cheap, decisions are expensive. If AI can generate something that compiles in seconds, the real question is whether we should build it at all.
I’m spending less time nitpicking implementations and more time asking why we’re choosing one path over another. I’m teaching the team to spot the decisions that are hard to undo and to slow down when they find one.
Context is part of the job
AI has no memory of your organization. Your team does. It doesn’t know why a decision was made three years ago. It doesn’t understand customer pain that never made it into a ticket, or the political and operational constraints that shape real-world software.
So helping engineers build and share context matters more than it used to. That means writing down why we made a call and pairing engineers with product and design earlier than feels comfortable.
Redefining growth
There’s a lot of anxiety around career progression right now. Writing harder code used to be the obvious path to senior, and that path isn’t as clear anymore. I don’t have a perfect answer. What I’ve been telling my team is that growth now looks like better judgment and knowing when not to automate, and none of that throws away the craft people worked hard to build.
What the job is now
I don’t think engineering leadership has become easier. I think it’s become more exposed. Tools take a lot of the friction away, and teams still need help agreeing on what they’re actually trying to do. They still need mentorship and trust, even if the shape of that work keeps changing.
Lately I’ve been orienting my team less around output and more around clear goals and ownership. The better we get at those, the more the tools actually help.
I’m still figuring this out. Some days it feels like I’m letting go of things I used to be really good at. Other days it feels like I’m finally seeing the job for what it always was. The code changed, but the people didn’t. That’s still the job.