Why I Built This

You know the hunt is a second job. You do your day, then you clock into the night shift: forty tabs open, every one of them wanting a tailored resume, a fresh cover letter, its own answers to the same questions worn slightly different each time. And you can't even tell which of the forty is worth an evening until you've already spent the evening.

So you reach for the AI. Everyone swore it would lift the weight. Here is what it actually does. Ask it the easy way and it hands you something generic and confident, a cover letter with an achievement you never earned bolted on to cover a gap, the kind of thing a recruiter smells in a second. Ask it the right way, feed it everything, make it research the company, weigh the role, get it genuinely good, and you find you have signed up for a third job: the same climb, start to finish, for every single posting. Lift became a haul. The thing you brought in to carry the second job just handed you a third.

I know how that feels. I lived it.

Let me be honest about who was sitting there, though, because it matters. I'm a builder. What I really wanted was to build something real, proper infrastructure, an AI team that actually holds together instead of falling over the moment real life touches it. And you don't build that against a toy. You need a real problem, a hard one, ideally one that's yours so you can't fake your way past the messy parts.

I also needed a job. I'd left one, and I'm particular about the next. I wasn't chasing any title with a salary attached, I was looking for the right next challenge, which means most of the work is deciding, not applying. Is this one even worth my time. That's the real labor of a senior search, and it's the part everyone tells you to skip.

So I had a problem worth building against and a build looking for a problem. I pointed one at the other.

And I did it properly. I set myself up in Claude, and not lazily. I spent a long session on each chapter of my career, a few real pages each, the actual story, the portfolio, the references, then a scoring doc for what I actually wanted in a role. So it knew me, top to bottom. Knowing me was never the issue.

Every single job was the issue. Each new opportunity meant a fresh session, and every session ran the same way. Deep research on the employer first, and I mean deep, the kind that chews through tokens like nothing else. Then score the role against my criteria. Then build the thing, ready to send. Two hours later I had one application worth its salt, and the session was burned to the ground. Next job, open a new one, do the whole thing again. And underneath all of it, try to keep track of everything I'd already chased. That was what drove me up the wall. Not that it forgot me. That every job cost me a full session, and there are a lot of jobs.

Which meant I knew exactly what to build, because I'd been doing it by hand for weeks. Automate the work. Never the decision. The deep employer research. The scoring against what I actually want, so I know in a breath which roles are worth my evening and which to let go. The drafting. The tracking, all of it in one place instead of scattered across dead sessions. Do the heavy lifting once, not once per posting.

And one line I was never going to cross: it does not apply for me. Pointing a machine at a job board to fire off applications while you sleep is exactly how good people drown in everyone's spray, how a real candidate turns into one more piece of noise in a pile no one reads. Send by hand, or don't send. The judgment stays mine. The send stays mine. That was never up for discussion.

So I built it as the real thing I'd wanted to build all along. Not a weekend wrapper around somebody's API. I come from product and design, and I've spent this whole stretch deep in how you get an AI team to behave instead of handing you confident nonsense. So instead of a pile of static docs, I gave it real, connected knowledge: a structured store of what's true about you, that every score and every line draws from and nothing it writes ever wanders past. If it only speaks from what's real, it can't invent.

In the product you never see that as a graph or a database you poke at. You see chapters. Your career breaks into chapters, and each one gets an expert who goes deep on that stretch of your life. They ask you the things a resume never had room for. Together they're a team that knows your whole story, and they only ever know what you've confirmed. There is nothing for them to make up.

For a while it was just mine. A tool built for one person, the one running my exact search. That was the whole ambition.

Then I showed it to a few friends. The reaction was the kind you can't fake. This is great, they said. You have to put this out. People need this.

So I put everything else down and made it real.

Scored

Stripe

Senior Infrastructure Engineer

strong fit

84

Build the apply kit

So here it is. It does what you'd do on your best day, the day you had unlimited hours: research a company properly, weigh whether a role is genuinely worth your time before you sink the time, and when it clears the bar, build the application from your truth and nothing else. Then it stops and hands you the send. Always you.

I built it for me first. That is the whole reason it might work for you. It was made by someone who sat exactly where you're sitting, who was picky about what came next, and who refused to ship the hopeful version.

If you want to see how I build things like this, that lives over on skakeny, my other home on the web. If you just want to start, the studio is right here.