Computer Vision Engineer at Boneprox

PVH (Tommy Hilfiger/Calvin Klein); Göteborgs kommun; On-site; SEK 650,000 - 800,000; Full time

Job summary

Boneprox is seeking a Computer Vision Engineer to work hands-on with real clinical data and production‑grade computer vision models. You’ll train, evaluate, and improve detection, segmentation, and keypoint models on dental radiographs, collaborating with a small, technically driven team.

You’ll bridge data quality, annotation pipelines, and backend integration to move prototypes into reliable production features, contributing across the stack from modeling to deployment in a clinic‑focused SaaS

Job description

Computer Vision Engineer at Boneprox

Introduction Do you want to work hands‑on with computer vision in a product that’s actually used in clinics, while growing broadly as an engineer? At Boneprox, you’ll build and refine the vision models behind our diagnostics, working closely with the rest of a small, technically driven team.

About us

Boneprox develops SaaS solutions for the dental industry, with a focus on diagnostics and more efficient workflows. Our products are used by clinics across Europe and combine modern web technologies with advanced AI. All development is done in‑house from our office at Järntorget in Gothenburg. The team has built the product from the ground up, and several members have been there since the beginning. Today, we are a small organization of around eleven people across three companies, where the development team works closely together in a flat and technically driven environment.

The role

This is a role for someone who wants to get hands‑on with computer vision on real clinical data, not toy datasets, not research demos.

You’ll work on training and improving detection, segmentation, and keypoint models for dental radiograph analysis, digging into annotation quality and dataset structure, running experiments, and iterating on model performance. You’ll work closely with our annotation pipeline and QA process to make sure the data feeding the models is actually good, and you’ll help push predictions from “works in a notebook” to “works reliably in production.”

The role isn’t isolated to modeling either. You’ll work with the backend team on getting models served correctly, and contribute to how new diagnostic features get evaluated and shipped.