Working Student, Computer Vision & Fullstack Engineer
Location: Hybrid / Munich
Hours: ~20h/week (working student status)
Start: ASAP
About the role
We build AI solutions for engineering, applying modern computer vision and ML to real engineering problems. Youll work across the full slice: training models, building the backend services that serve them, and shipping the interfaces engineers actually use.
What youll work on
* Training and fine-tuning detection, OCR, and segmentation models on real and synthetic data
* Designing synthetic data and augmentation pipelines that transfer to production
* Classical CV preprocessing as fast, reliable baselines. Not everything needs a GPU
* Backend inference services and data pipelines in Python
* Packaging, deploying, and operating what you build
Must-have:
* Strong CS background, BSc completed from a reputable university. MSc in progress is welcome
* High ownership. You take problems end-to-end, ship without hand-holding, and own what you build in production
Computer Vision:
* Solid grasp of modern deep-learning vision: detectors, training loops, losses, evaluation, augmentation, overfitting and regularization
* Comfortable with PyTorch
* Working knowledge of classical / legacy CV with OpenCV. You should reach for these before throwing a GPU at every problem
Fullstack & DevOps:
* Hands-on fullstack experience, frontend and backend, in real projects
* Have built and deployed something end-to-end, not just localhost
* Docker, Linux shell, SSH, bash scripting
* CI/CD experience (GitHub Actions or similar)
Nice-to-have:
* Cloud-native: Azure, AWS or GCP, managed databases, object storage, IaC
* 3D vision / geometric deep learning: B-rep, STEP/IGES, OpenCASCADE, GNNs on CAD topology, point clouds, mesh processing
* Foundation models for vision (SAM, GroundingDINO, VLM finetuning)
* Interest in AI for engineering: CAD, manufacturing, technical documentation
What we offer:
* Strong compensation, above-market rates for working students
* Top-tier AI tools and budget covered
* Fast-paced environment. Small team, short feedback loops, your code ships the same week
* Flexibility. Remote-friendly, hours that work around your studies