I co-founded bydhome as an equity partner, which is a polite way of saying I’m the one who gets the message when something breaks at 11pm. Building it from nothing changed how I think about what “AI engineering” even means.
The AI is the easy part
The models and pipelines — turning messy, human-entered data into something structured and searchable — were the part I was most excited about, and honestly the smallest slice of the work. The rest was:
- Schema and data model — get this wrong early and everything downstream fights you.
- The boring plumbing — auth, infrastructure, the pipes nobody demos.
- Shipping incrementally — real users, so no big-bang launch; every change has to be safe.
Owning the outcome
There’s a difference between “I built a model” and “I own whether this product works.” As a co-founder I can’t hand off the hard 80%. That constraint made me a better engineer than any tutorial did.
Being honest about status
The core marketplace is live; the seller dashboard is still in progress. I’d rather say that plainly than pretend it’s finished — and it’s exactly how I’d talk about the state of things on your team.
Why I write this down
Because the thing that makes me worth hiring isn’t that I can call an API. It’s that I’ve taken something from a hard problem all the way to real people using it, and stuck around to keep it running.