Software I delivered ten years ago is still in production — and those clients still have my email. CTO of a ten-person build, embedded fractional CTO, one-man team: architecture to production to a handover your team can live with.
A short reply to see if I'm the right fit. I won't waste our time.
01 / Fit
I take that seat: architecture, roadmap, priorities, and the call on what's worth building. AI tooling made prototypes cheap, so that question is often settled by spiking it and testing — not by weeks of planning.
I've shipped a production AI assistant that books, reschedules, and takes payment — and can't claim anything the system didn't confirm. The engineering around the model is the product; I've built it.
Hand me a scoped problem or an unscoped one. I've delivered six-month integrations on my own and led ten-person builds. If you run an agency, I take overflow work under your name.
The people who built it moved on, or it grew faster than it was designed to. I make it stable, understandable, and easy to extend — usually without a rewrite.
02 / Practice
Not just the code. The data model, the infrastructure, the third-party mess, the stakeholder conversation. One accountable person from “we need this” to “it's running.”
AI made building cheap — the quickest answer to “should we build this?” is often a working spike. But data, payments, and migrations don't forgive the way code does. Knowing which is which is most of the job, and it's how a prototype avoids quietly becoming production.
On my largest build, the client listed dozens of hard rules. In practice, nearly every one needed exceptions. Now I watch how people actually work before locking anything in.
I build systems feature-complete, automated, and documented — so when the work is done, you decide: hand it to an in-house maintainer, or keep me as long-term support. I've done both. On my largest build I trained a junior team to take over at a fraction of the cost; my longest-running client keeps me on years later.
03 / Engagements
What was needed, the decisions that mattered, and what I'd do differently now. No invented metrics.
Built around the movers' actual process, not a vendor's idea of it. Three years leading the build, then a handover to a junior team I trained.
An assistant that can't invent bookings: it only acts through the same validated APIs as the app. Now multi-tenant, and I'm still with them.
Made the ATS and the scheduling platform finally share one truth. Shipped solo in six months, plus a roadmap the shareholders could act on.
The vendor they depended on shut down. I built the replacement they own — six people run the entire company on it.
04 / Plan
From first message to scoped work — nothing hidden, nothing performative.
What you're building or what's in your way. I reply with questions and a fit check.
What I'd do, what it costs, what I'd skip — in writing. Wrong fit? I'll say so.
Scope it, ship it stable, hand it over. Your team runs it — or I stay on.
05 / FAQ
06 / About
I'm Aleksandar, an independent software engineer in Belgrade. Small teams are my natural habitat — enough ownership to do things properly, no layers between the problem and the person solving it. Laravel is my default backend, Expo for mobile, but the problem decides the stack. Off the clock I build my own tools — HTML-first, open source when they're ready — at softkittens.com, including effortless.run, my framework.
A short reply to see if I'm the right fit. I won't waste our time.