Senior Computer Vision & Backend Engineer (Python) – Remote

Turn a phone camera into a physiotherapist.Breathment is a Munich-based digital health startup. We digitize pulmonary rehabilitation for people living with chronic respiratory diseases such as COPD: our app guides patients through therapy exercises at home, watches their movement through the phone camera and gives feedback in real time, while physiotherapists and physicians follow their progress remotely through our telemonitoring platform. We work with German clinics, physiotherapists and health insurers, and every release reaches real patients.We are looking for a senior engineer who is at home in both worlds: computer vision and production backend. You will own the path from neural-network output to what the patient sees on screen and what the therapist reads in the report.What you will buildOwn the computer-vision pipeline that turns pose-estimation output from our neural networks into repetition counts, posture feedback and set-by-set exercise analysis.Design the reporting layer that converts raw movement data into clinical reports physiotherapists act on.Build and extend our Python backend (FastAPI, MongoDB, Redis): the API behind the patient app and the clinician platform.Ship to production on Kubernetes with GPU inference services, and stand behind what you ship: tests, monitoring and clean handovers.Analyse requirements together with product, physiotherapists and the mobile team, and propose solutions instead of waiting for tickets.What you bring5+ years of professional software development in Python, including at least a year running a FastAPI (or comparable async) service in production.Hands-on experience with deep learning frameworks (PyTorch or TensorFlow). You have built real product features on top of model output, not only trained models in notebooks.Solid experience with NoSQL databases, ideally MongoDB, and with data models that survive growth.A university degree in computer science, electrical engineering, mathematics, physics or a related field from one of the universities listed under "Education" below.A fluent Git and GitHub workflow: branches, reviews, CI.Strong analytical thinking and an independent, structured way of working. You can take an ambiguous problem and drive it to a shipped result in a remote team.Professional working proficiency in English. German is a plus, not a requirement.Genuine interest in startups: you want responsibility, and you want to see the effect of your work on patients.Nice to havePose estimation, human motion analysis or video processing (OpenCV, MediaPipe or similar).Docker, Kubernetes and cloud deployments; workflow engines such as Temporal.Daily use of AI coding tools with a clear habit of verifying what they produce.Software in a regulated environment: medical devices, health data, GDPR.Why BreathmentImpact you can see: your code helps patients rehabilitate at home instead of waiting months for a clinic place.Remote-first: work from anywhere.A small senior team with short decision paths and direct access to founders, medical team and users.A modern stack: Python 3, FastAPI, MongoDB, Redis, PyTorch, Kubernetes, GPU inference.EducationFor this opening we are recruiting graduates of the universities that the Turkish Council of Higher Education (YÖK) designates as Research Universities (Araştırma Üniversiteleri), full members as published on YÖK's website at the time of application. Among them: Istanbul Technical University, Middle East Technical University, Boğaziçi University, Koç University, Bilkent University, Sabancı University, Hacettepe University, Yildiz Technical University, Istanbul University, Ankara University, Ege University, Gebze Technical University and İzmir Institute of Technology. If your degree is from elsewhere, this particular role is not a match, and we would rather tell you now than after you have spent time applying.How we hireApply on LinkedIn with your CV and answer the short screening questions. Please read them carefully; they decide whether we see your application.If you pass, we send you six written questions by LinkedIn message (about 20 minutes, no cover letter). We read every answer ourselves and prefer specific over polished.30-minute intro callTechnical interview including a coding challengeFounder conversation and offerWe aim to complete the whole process within one month.