At Starsky, the Perception team is responsible for processing sensor information and making it available to the other teams in a clean and consistent format. The models and algorithms developed aim to achieve robust real time detection and tracking of our truck and other objects in the local environment, including lane lines, vehicles, and pedestrians.
As a Robotics Perception Engineer, you will be responsible for filtering, fusing and post-processing the outputs of different deep learning models and sensors. You will apply state of the art tracking and fusion algorithms which are robust to sensor noise and environmental variability. This requires working with teams across the driving stack to scope requirements and understand the strengths and limitations of different modules.
We are looking for an individual who can bridge the gap between a research project and a functional product safely and efficiently. Candidates should be able to approach the problem with first principles and work within design constraints.
Essential
2+ years experience in perception system of mobile robots
2+ years experience in C++/Python development in a fast paced production environment
Background in linear algebra, probability, 3D geometry
Experience with some of the following: multi-camera sensor fusion and camera-radar sensor fusion, real time tracking of objects and lanes, real time mapping and localization, camera and radar sensor calibration
Experience in developing on linux environment
Nice to have or willing to learn
Experience with ROS
Experience in sensor noise analysis
Experience in machine learning/deep learning
Starsky Robotics
