Why I Left a Career in Engineering to Build AI for AEC

I grew up in a filing cabinet

My father's engineering firm is forty years old. I grew up in it the way some kids grow up in a restaurant, doing homework in the conference room, learning what a boring log was before I learned long division, watching the same people stay late on the same kinds of nights.

The thing I remember most is not the engineering. It is the other work. My father, one of the best geotechnical engineers I have ever known, spending a Sunday rewriting proposal boilerplate I had watched him rewrite a dozen times. A senior associate digging through binders for a project from 2003 because she was sure we had drilled near that site before and only she would know. Forty years of reports in a room, and the only working index to them was the people who had written them.

I did not have words for it then. I just knew that the people who were best at the actual work spent an enormous amount of their lives not doing it.

I became an engineer, and saw it from the inside

I studied engineering and went to work, partly at the firm and partly away from it. And the pattern I had felt as a kid turned out to be a law. The most skilled, most expensive, most irreplaceable people in any firm were the ones losing their evenings to work that did not need them: assembling proposals, transcribing field notes at 7 p.m., scoping every estimate by hand because the judgment lived in one head, answering questions the archive already contained.

It was not a small-firm problem. It was not my father being disorganized. Every firm I saw ran the same way, and everyone treated it as simply the cost of doing the work. The senior bottleneck was load-bearing. Nobody questioned it.

Then I understood why nobody had fixed it

When I finally looked up from the inside, the numbers explained the feeling. Construction and the broader AEC industry, something like twelve trillion dollars globally, is one of the least digitized major industries in the world. While finance, healthcare, and logistics were being rebuilt around software, our industry was still running on PDFs, shared drives, and tribal memory.

The AI gap is even starker. Only about 27% of AEC professionals use AI in their work, against roughly 72% of organizations across all industries. And the reason is not money. As Bluebeam's CEO put it bluntly in early 2026, the biggest barriers to AEC technology adoption "aren't cost. They're complexity, culture, and connection." The tools that transformed other industries were never built for the way an engineering firm actually works: the proposal, the report, the estimate, the forty-year archive.

So firms like my father's were stuck with two bad options. Use generic AI that did not understand their workflow, or build their own, which they could not, because they are full of civil engineers, not machine-learning engineers. The expertise to build and run these systems is exactly the expertise an AEC firm does not have in the building. The right tool for firms like ours simply did not exist, and the people who needed it most were the least equipped to make it.

Why I left

I left a career in engineering because I could not stop seeing it. Not the technology gap as an abstraction: my father's Sunday. The associate and her binders. The junior engineer who should never have had to interrupt a principal to find out whether we had seen a site before.

I did not leave to replace engineers with AI. That framing has always struck me as both wrong and a little insulting. The entire value of a firm like ours is the judgment of its senior people: what a soil profile means, what a project really requires, which risk is worth pricing in. You do not want a machine making those calls. You want those people to have more time and energy to make them.

What I wanted to build was the opposite of replacement. A connected layer that meets engineers inside the work they already do, that drafts the proposal from the firm's own history, turns field notes into a report, makes forty years of archive searchable in seconds, gives the whole team structured access to what the principal already knows. So the senior people only ever touch the part that genuinely needs them.

What I am actually after

There is a version of AI in our industry that I do not want to build, one that floods firms with another disconnected app and calls it transformation. The opportunity I care about is quieter and, I think, much bigger: take the work that is throttling the most capable people in every firm, and lift it off them, in the tools and workflows they already live in.

McKinsey has estimated that technology and innovation could raise construction productivity by 50 to 60 percent. I do not think that number gets unlocked by a flashy model. I think it gets unlocked one firm at a time, by giving the people who already know how to do the work back the hours they are losing to everything around it.

For the firm I grew up in

I am building this for my father's firm, and for the thousands like it, the regional, multi-generational practices that hold an extraordinary amount of hard-won knowledge and almost no way to use it at scale. They are not behind because they are not smart. They are behind because no one built the right thing for them, and they could not build it themselves.

That is the gap I left engineering to close. Not to take the engineers out of the work, to give them back the part of it they fell in love with, and to make sure the forty years in that filing cabinet outlive the people who put it there.