Calibration Team Lead for Autonomous Transport

To ensure high-quality and stable operation of autonomous vehicle recognition and localization systems, it is essential to have good sensor data. It is crucial to calculate with high accuracy the relative positions of LiDARs, cameras, radars, accelerometers, and gyroscopes relative to each other and relative to the vehicle body, as well as to determine the internal properties of individual sensors, such as camera lens distortion parameters. The development of sensor calibration procedures during assembly and maintenance, along with mechanisms to maintain calibration status during operation, is the focus of the Calibration Development team.

What tasks await you

Creating mechanisms for in-motion sensor recalibration This includes: adapting existing "heavy" offline algorithms to hardware limitations during the autonomous vehicle pipeline's operation. Analyzing and implementing SOTA solutions in the field of online calibration. Quickly determining the loss of calibration status.

Improving calibration algorithm accuracy, reliability, and performance Specifically: significantly (multiple-fold) improving algorithm accuracy. Diagnosing and systematically fixing unsuccessful calibration results. Profiling and refactoring algorithms to improve their performance. Utilizing GPU for computational acceleration.

Automating calibration procedures This includes: eliminating human factors from calibration acceptance. Simplifying calibration procedures to enable their execution by unqualified personnel. Reducing the dependency of procedures on specific execution and verification conditions.

We expect that you

• Have managed a team
• Are proficient in C++
• Know how to improve code performance and optimize resource consumption
• Have a solid mathematical background and are ready to dive into the world of SOTA calibration solutions
• Know classical algorithms and data structures

Will be a plus

• Have experience working with sensor data
• Have worked on solving SLAM problems or optimization on factor graphs
• Have experience with CERES, G2O, or GTSAM
• Have used GPU for computational acceleration
• Have knowledge in the field of computer vision