Kodiak Staff Autonomy Software Engineer role on the Calibration, Localization and Mapping team. Lead design and development of sensor fusion and pose estimation systems.
Responsibilities
Set the technical strategy and architecture for our onboard navigation and pose estimation systems.
Lead the design and development of advanced sensor fusion and pose estimation approaches across IMU, camera, and lidar sensors, anticipating failure modes before they arise in the field.
Drive the direction of visual and lidar odometry algorithms, including evaluating build-vs-adopt tradeoffs for new techniques.
Define and evangelize navigation metrics and evaluation methodology used across the org to assess algorithm and system performance.
Serve as a technical escalation point for complex, ambiguous, or cross-cutting localization problems.
Evaluate sensor performance and optimize sensor selection.
Verify correct sensor calibrations, settings, and uncertainty characterizations.
What you’ll bring:
Strong technical background. BS, MS or PhD with academic or industry experience in building navigation, localization, and odometry systems.
Expert in C++ and strong problem-solving skills.
Strong software engineering skills and experience building high-reliability deployed applications . You write well-designed, highly maintainable code. Code is your craft.
Deep experience transforming raw sensor data from GNSS, IMU, wheel speed, etc. into 6-DOF pose using Kalman Filtering, factor graphs, or comparable state estimation techniques — and judgment on when each approach is appropriate.
Qualification
BSExpert in C++Strong software engineering skillsExperience with computer visionHistory of mentoring engineers
Required
BS, MS or PhD with academic or industry experience in navigation, localization, and odometry systems
Expert in C++
Strong software engineering skills
Experience building high-reliability deployed applications
Deep experience transforming raw sensor data into 6-DOF pose using state estimation techniques
Experience with computer vision, visual odometry, and/or odometry from point clouds