Reesure is AI-native rent operations: our agents run the full cycle for landlords and property managers — from rent due, to payment missed, to arrears case resolved.
Every month across Europe millions of rent payments go out late, and behind each one sits someone doing the same grind by hand: checking who paid, writing the reminder, deciding between a friendly nudge and a formal notice, arranging a payment plan, keeping it legally defensible, explaining it to the owner at month end. High-volume, high-stakes, almost entirely manual — a near-perfect problem for agents. Ours watch every lease, open a case when a payment is late, decide the next step within the rules of that jurisdiction, write the message in the right tone through the right channel, and escalate only when they should. Not a chatbot bolted onto a dashboard: agents that carry a case from “payment missed” to “resolved”, with humans on the decisions that matter.
Which makes it hard in the interesting way. An agent that improvises in a regulated debt-recovery conversation is a liability, so it all sits on a deterministic layer underneath — guardrails, statutory windows, evals, audit trails. You’re not prompting your way to a demo; you’re building agents that are allowed to act on someone’s money and have to be right. That’s how we work internally too: the whole team directs AI to write the code and spends its judgment on evaluating the output.
What you’d actually build
Most of what we build doesn’t have a screen. The weight of the product sits in backend processes that run unattended: scheduled jobs that check every lease every day, sync engines talking to external systems, reconciliation logic matching money to leases, the recovery engine deciding what happens next on hundreds of open cases. Nobody is watching when they run, so they have to be right — and that’s exactly where we need reinforcement. The team is small: our Technical Lead owns architecture and delivery, and one senior builder owns UX and product quality. What we’re short of is someone senior on the process and engine side.
So the role is full-stack, but clearly backend-leaning. Concrete examples of what the team has shipped and is shipping:
Integrations with property management systems — We import entire portfolios from a PMS — owners, buildings, units, leases, payment history — and keep them in daily sync. That means reconciling two systems, handling contracts that end on one side but not the other, and making an import of 2,000 units something a customer can complete in an afternoon.
Payment plans — A tenant can’t pay €1,800 at once but can pay it over four months. That’s an instalment schedule, compliance tracking (did they keep to it?), automatic recalculation when they miss one, and a clear picture for the landlord of who is actually recovering and who isn’t.
Arrears and recovery management — When rent is late, a case opens and runs a protocol: reminders through the right channel, escalating over time, respecting statutory notice periods and local regulation. Some of it is AI-driven communication, all of it needs guardrails.
Payments infrastructure — Payment links, direct debit mandates, reconciling incoming payments to the right lease and period, retry logic when a collection fails, multi-currency, Stripe Connect across many payout accounts.
Reporting owners can trust — Monthly statements, reconciliation exports for accounting systems, annual rent indexation runs across a whole portfolio.
What we’re looking for
Solid backend experience — Ideally C# / .NET. If you’re strong in another typed backend language and want to move to .NET, we’re open to that.
Comfortable across the stack — Most of your work is backend, but you’ll touch React / Next.js when a feature needs a front end. You don’t have to be a design specialist; you do have to care what the user ends up with.
You think in flows and edge cases, not just endpoints — Our domain is full of “but what if the tenant pays half, two days late, in a month where the rent was indexed?” We need someone who spots that before it reaches production.
At least 3 years of hands-on experience working with AI, and real range across it — A hard requirement, not a nice-to-have. Cursor and Claude Code as daily drivers — past “I’ve tried it” and into having opinions. Agentic tooling strongly preferred (OpenClaw / Hermes or comparable frameworks), where agents do multi-step work rather than answer single questions. And breadth across models beyond OpenAI, Anthropic, Google and xAI: open-weight and self-hosted (Llama, Mistral, Qwen, DeepSeek) and platforms like Together, Groq, Fireworks, Ollama or vLLM. We want someone who has compared models on a real task and can say which to use where, on quality, latency and cost.
The judgment to evaluate what AI produces — The whole team directs AI to produce code; the skill we hire for is what happens next — catching what the model got wrong, where it introduced complexity, where it quietly broke an assumption. Writing code is cheap now; knowing which output to trust isn’t.
Ownership — Small company, real customers, real money moving. On the things you own, you follow through — including checking after release whether it actually did what it was supposed to.
Fluent English — Written and spoken. Our team is Dutch and Macedonian; everything technical happens in English.
Nice to have: experience with payments (Stripe, PSD2, direct debit), fintech or another regulated domain, having built evals or guardrails for AI features in production, or having worked in a startup where you had to make the call yourself.
How we work
Design → Build → Test → Demo — Small scopes go end-to-end with one person; larger ones split design up front and testing at the end. Everything gets reviewed.
Daily stand-up, sprint planning, demo day — Not heavy process, but a rhythm you can count on.
You’ll work directly with the founders and the Technical Lead — No layers, no ticket-shuffling. If you think we’re building the wrong thing, say so in planning — that’s expected, not tolerated.
Practical
You live in Skopje, or you’re moving there — We have an office in central Skopje and we want people in it — on-site several days a week. This is not a remote-from-anywhere role, and we’re not able to sponsor relocation from outside the region at this stage.
Full-time, long-term. Independent contractor arrangement with Reesure Operations B.V. (Netherlands).
Compensation depends on experience — we pay competitively for the Skopje market and we’ll be straightforward about the range in the first conversation.
How to apply
Send a short message with your CV or GitHub. Don’t write us a cover letter. Instead, tell us two things in a few sentences each:
A piece of software you built where the logic was genuinely tricky — and what you got wrong the first time.
Your AI setup: which tools and models you actually use, for what, and one case where you picked a model or approach other than the obvious one and it paid off.
Our process
Short: one intro call, one structured technical challenge, one conversation about the challenge, decision. No five-round interview loops.
Sound like you?
Send us your CV or GitHub and the two short answers above. We read everything ourselves.