Gideon Brothers develops autonomous, collaborative and modular robot platforms and autonomy technologies for unstructured, indoor, human environments. Our robots are designed to assist operations in various industries. In our offices in Zagreb and Osijek, we have gathered a team of 70+ super-talented people, including 9 PhDs and 35+ Masters of robotics hardware and software engineering and related disciplines. Now we are looking for a Computer Vision SLAM Engineer for our Zagreb office!
Scope of work:
In this position, you will primarily be responsible for developing parts of the SLAM algorithm based on a stereo vision system. SLAM enables our robot to map the environment and then localize within the map without introducing any artificial changes to the environment. We are looking for a person who will work on both front-end and back-end SLAM components.
Develop and implement state-of-the-art SLAM algorithm based on a stereo vision system.
Improve the capabilities of stereo visual odometry, loop closing detection, and global and local pose graph optimization.
Improve and develop a real-time dense mapping algorithm and its integration with semantic information.
Work on extrinsic camera calibration and fusion with inertial sensors.
Work with unique datasets from extensive full-scale robot operation as well as acclaimed open-source datasets.
Identify technical limitations and devise strategies to overcome them.
Share knowledge by clearly articulating ideas through papers and presentations to technical staff and management.
Good understanding of linear algebra, optimization, numerical methods and other basic algorithms.
Good understanding of extracting image features and feature descriptors matching.
Strong C/C++ skills.
Strong engineering practices, coding by standards and good debugging skills.
Familiarity with Linux and docker environment.
Experience with computer vision libraries such as OpenCV.
Experience with parallel computing libraries CUDA/OpenCL.
Experience with camera calibration.
Good understanding of epipolar geometry and stereo matching principles.
Experience using Ceres solver and protobuff.
Experience with the development of visual odometry, simultaneous localization and mapping (SLAM), scene reconstruction, state estimation and fusion with inertial sensors.
Experience with real-time dense mapping using stereo cameras.
MSc or PhD in Computer Science, Robotics or related field.