# Backend Engineer (Integrations & Applied AI; Mid-level) at Relu

> Location unclear · Mid-level · Backend · open. Source: https://alderhq.app/jobs/personio%3Arelu%3A2840880 (read 2026-11-09).

## Facts (observed)

- Company: [Relu](https://alderhq.app/companies/relu)
- Location text: Leuven
- Where it can be done from (inferred): Location unclear (evidence: "Location: Leuven")
- Posted: 2026-10-13 (27 days ago)
- First seen by alder: 2026-11-04 · last verified: 2026-11-08
- Reposted: 0 times · edited: 0 times
- Apply on the employer's site: https://relu.jobs.personio.de/job/2840880
- Skills listed (inferred): Python, LLMs

## Pay and sponsorship

- Posted pay: not posted

## Description (the employer's text)

### Your mission

We're looking for a Backend Engineer, Integrations & Applied AI to join our Automate Platform team. You'd own the inbox pipeline from the moment a case arrives: get data in from every scanner and folder, understand the case, act on it, and keep the dentist in the loop. What you would work on:

- Integrations in: scanner platforms (3Shape, Medit, iTero, Shining3D, AlliedStar, DS Core, Dexis and new ones), folder sync, email, WeTransfer.
- Integrations out: LMSs (Labtrac, DentiSky and others) for pulling order data and pushing status, CAM and printers (Shapeware, Formlabs, SprintRay, Carbon), folder sync out, exports to 3Shape and Exocad.
- Automatic case entry: enter scanner data into the LMS. For LMSs without an API, a browser or computer-use automation.
- Rx digitizer: read prescription PDFs and scanned forms into structured data (teeth, shades, materials, due dates), across languages and lab abbreviations.
- Rx intelligence: from the Rx, notes, doctor and the lab's history, predict which CAD orders to create, and learn from what users accept or correct.
- Rx QC: flag missing, inconsistent or contradictory prescriptions.
- Scan Check in the inbox (backend): pick checks by order type, set blocker versus warning. The AI team builds the checks.
- Dentist messaging: per-case threads through scanner platforms, email and WhatsApp.
- Ask Relu: tools that act in the app, the knowledge base, model choice and cost, answer-quality evals.
### First 6 months

- Month 1: take over inbox poll health and two or three vendor integrations. Fix the known evidence leaks in the sync templates.
- Months 2 to 3: build the Rx extraction eval set from real cases. Ship the digitizer and normalization v2 behind a confidence threshold.
- Months 3 to 5: first LMS integration (pull data, push status), then automatic case entry for one LMS.
- Months 5 to 6: Rx intelligence v1, predicting CAD orders per lab and learning from corrections.
### Your profile

- Product ownership and vision: takes a raw user problem ("a lab says cases are missing", "order entry takes too long"), finds the real need, and scopes what is essential versus nice to have before building.
- 3+ years of backend engineering, strong Python.
- Has built many third-party integrations and handled OAuth and token refresh, rate limits, idempotency, retries, webhooks versus polling, and schema drift.
- Has shipped LLM features to production with evals: datasets, metrics, regression checks when prompts or models change, cost tracking.
- Owns problems end to end, from "a lab says cases are missing" to root cause, fix and monitoring.
- Pragmatic: picks a rules catalog with an LLM fallback over an elaborate ML system when that works.
Nice-to-haves

- Healthcare, dental or regulated-data experience (HIPAA, GDPR).
- Document extraction or OCR, or browser and computer-use automation.
- Workflow engines (ActivePieces, Temporal, n8n) or ETL pipelines.
- Dutch or French.
### Why us?

- A mission that matters: we're improving dental care and making higher-quality dentistry accessible to everyone, globally
- A fast-growing company where your growth goes hand-in-hand with ours — real impact, real ownership, early responsibility and skills in the technologies of the future
- The chance to work with the latest AI technologies and industry leaders in dentistry
- Fun team culture: after-works, bi-weekly sports, hackathons, retreats

---
Data: alder (https://alderhq.app). Cite https://alderhq.app/jobs/personio%3Arelu%3A2840880 and the date you read it. Observed facts come from the employer's own listing; location reach, level, family and skills are inferences.