Computer Vision Engineer, 3D Perception | Tier 1 VC-backed Physical AI | Germany (English-speaking)
AI Futures are partnered exclusively with a Tier 1 VC-backed Physical AI company on a Search for a Computer Vision Engineer, 3D Perception.
Founded by former research leads from Google and ETH Zürich, and backed by leading European and US deep tech investors, the company is building a perception foundation model that gives robots a real-time 3D understanding of the world around them. The initial application is autonomous mobile robots in warehouses and construction sites, with the long-term goal of a single perception stack that any robot, drone or vehicle can plug into.
The Role
- Design and train deep learning models for 3D perception: object detection, segmentation, depth estimation, tracking and scene understanding
- Work across multi-modal sensor data (RGB, depth, LiDAR, IMU) and build models that hold up in messy, real-world conditions
- Take models from research to deployment on edge hardware, optimising with TensorRT, ONNX or CUDA
- Build and scale data pipelines, auto-labelling and evaluation tooling across a fast-growing real-world dataset
- Shape the technical direction of the company as an early member of the perception team
The Candidate
- Strong Python and PyTorch, with hands-on experience training computer vision models
- Experience in at least one of: 3D vision, object detection, segmentation, SLAM, multi-object tracking, depth estimation or sensor fusion
- Experience working with real-world camera or sensor data, in industry, a research lab or a serious personal project
- MSc or PhD in Computer Science, Robotics, Electrical Engineering or a related field, or equivalent practical experience
- A builder mindset: you would rather see a model run on a real robot than chase another benchmark point
Nice to have
- Experience with vision transformers, foundation models (e.g. DINOv2, SAM) or neural scene representations (NeRF, Gaussian Splatting)
- Edge deployment on NVIDIA Jetson or similar
- C++ for real-time systems
- Publications at CVPR, ICCV, ECCV, NeurIPS or equivalent
English is the working language. German is not required. Remote working is possible with occaisonal on-site visits.
If you would like to explore this opportunity, please send your CV to will.childs@aifutures.group
AI Futures Ltd | Filling the AI skills gap ®