At Relu, our software turns dental scans into real dental appliances: crowns, night guards and more. We are looking for new team members to join our platform team, as a backend engineer, a front end engineer, or full stack.
Here is what you will find below.
Who are we looking for?
Open role
You would own the inbox pipeline from the moment a case arrives: pulling it out of whatever scanner platform, folder or mailbox it came from, working out what the case is, acting on it in the lab's own systems and keeping the dentist in the loop.
Five real tickets, unedited.
Open role
Our users are dental technicians. Many are not especially technical, and they spend most of the day in our software. Your job is to make it simple and to own how it looks and feels.
You would take a rough idea or a customer complaint to a prototype, test it with Thomas or a lab, and ship it. You would also review front end pull requests from the whole team, set the standards everyone follows and mentor Robbe.
Your first six months, roughly in that order.
Who would you work with?
Relu is about twenty people. Four of us are the platform team, which is where you would start. Tap a card to flip it and see what each of them works on.
What is the team working on?
We do not split into front end and back end squads. Everyone sees the whole system, and most people end up owning one or two of these areas.
Built in Vue 3 and TypeScript, with dense tables, upload queues, order states and approval flows. The test is whether a technician can run their whole bench on it without training.
Python, FastAPI, PostgreSQL, SQLAlchemy and Alembic, running queues, retries, state machines and billing. The order model has to stay correct when a lab sends four hundred cases before lunch.
We connect to 3Shape, iTero, Medit, Shining3D, AlliedStar, DS Core, Dexis, Dropbox and Google Drive. That means OAuth tokens that expire in creative ways, rate limits that arrive at the wrong moment and file formats with strong opinions.
A mesh viewer with editing tools, measurement and implant planning geometry. Scans with hundreds of thousands of triangles have to run smoothly on a lab laptop.
They pull scans out of email and cloud drives, read the prescription, open the right order, chase the missing file and answer the customer. They run in production for real labs today, and we build with AI ourselves: Claude Code and Codex are part of how we ship.
CI/CD, environments, cost and monitoring with Grafana and Loki, plus the part most startups skip: the change control and audit trail a medical device requires.
More information
Lui grinds their teeth at night and wakes up aching.
A two minute scan of Lui's mouth. No tray, no impression goo.
The scan is opened in Relu and designed in Relu, a technician checks and prints.
The guard is fitted.
Lui sleeps through the night.
No brochure version here, just the week you would be joining.
The team gathers on the sofas before the day starts, and we are running out of couch space.
Code you write this sprint is in front of a real lab by the next one.
Most days, most of the company is there: thirty minutes eating, thirty minutes of something else.
Or a nap, or whatever gets you back to focused work afterwards. There is an internal league where you can collect relo points.
Every quarter you get a whole day for whatever you have been meaning to look into, as long as it makes us better: new prompting techniques, how AI models compare, a UI that adapts to the person using it, or what makes a platform easier to trust. You choose the topic and show the team what you found.
People are often surprised by how much they get to decide here. There is no architecture committee and no quarterly roadmap to petition. You pick how the thing you own should work, and you live with it.
That is the trade at a company this size: you get the whole surface and the responsibility that comes with it, earlier than you would anywhere bigger.
We built the single inbox in two months and the agent platform in three. What you write reaches customers in the next release, about every two weeks.
That keeps the gap between making a decision and finding out whether you were right short, which is the main thing that makes this work fun.
The models are ours. We trained them on dental data most companies cannot get near, inside a regulated medical device, which is years of work a competitor cannot shortcut.
Dental labs use it every day. It makes good dental work cheaper to produce, which in the end means more people can afford it.
The part that is hard to describe is shipping something and hearing from a lab that it helped, which lands differently than a sprint burndown.
Every company says this, so here is the version worth anything: you can check. Relu is twenty people and you will meet most of them before you sign anything. Ask what they work on, what they like and what we should improve. Their answers will tell you more than this paragraph can.
There are also workaways, watch parties and the latest equipment, and people are trusted to manage their own focus, whether that takes a nap or a walk.
Relu is growing quickly: 80,000 cases a month and rising, with a team still small enough for one person to shape how things work. Join now and your scope and responsibility grow as the company does.
We are a small team building one connected system, and the fastest way to build it is still people in a room. We would rather tell you now than in round three.
Not at the moment. You need the right to work in Belgium already.
English. Dutch is useful but not required.
Two roles are open. Pick the one that sounds like you.
Backend Engineer Product Engineer
Not sure you fit? Apply anyway and we will find out together.
Questions first? Mail alexandra.ortmann@relu.ai.