3 months ago
Berlin, GermanySenior
Responsibilities
- Design and implement deep-learning models for 3D perception, including object detection, semantic segmentation, and occupancy prediction.
- Develop and optimize multimodal networks that fuse LiDAR, radar, and camera data for off-highway autonomous vehicles.
- Contribute to Vision-Language-Action models combining perception and language inputs for physical AI.
- Optimize training and inference pipelines for real-time deployment on NVIDIA edge GPUs.
- Lead perception data initiatives covering data pipelines, data curation, and model evaluation.
- Collaborate with interdisciplinary teams to integrate perception systems into the full autonomy stack.
Requirements
- Master’s or PhD in Computer Science, Robotics, Electrical Engineering, or a related field.
- Proficiency in Python and strong experience with PyTorch.
- Deep expertise in 3D perception and LiDAR-camera-radar sensor fusion.
- Practical experience deploying deep-learning models in real time on embedded hardware using TensorRT, ONNX, and Jetson/Orin.
- Solid understanding of machine learning, deep learning, and autonomous systems.
- Preferred: experience with transformer-based perception architectures or Vision-Language-Action models.
- Preferred: familiarity with BEV perception and multitask learning.
- Preferred: experience with C++, ROS, and mmdetection.
- Preferred: experience with perception in off-road, adverse-weather, or otherwise challenging conditions.
- Preferred: publications or significant industry experience in deep learning for autonomous driving or robotics.
Benefits
- Attractive compensation package and stock options.
- Beverages on-site and regular social events.
- Access to top-tier researchers, engineers, and thought leaders.
- Opportunity to influence robotic technologies and tackle significant technological challenges.
- Assistance with relocation to Berlin.
- Berlin/Potsdam-based role with an office environment.
Categories
ML EngineeringRobotics
About Sensmore
Sensmore builds AI-driven automation software and perception systems that turn heavy mining and construction machines (e.g., wheel loaders and LHDs) into autonomous robots, integrated with site safety, networking, and operational workflows. It sells a vertically integrated stack, including Material Handling Intelligence and 4D imaging radar, to operators and OEMs. Founded in 2022, the privately held company is headquartered in Berlin and Potsdam and is backed by Point Nine Capital and public EU programs.