Kodiak Staff Autonomy Perception Engineer leading technical direction for learned scene estimation and multi-object tracking in autonomous vehicles. Develops and deploys state estimation and sensor fusion algorithms using AI models.
Responsibilities
Define and lead the technical direction for learned scene estimation within the Kodiak Driver
Architect and implement principled approaches for fusing information across multiple modalities: camera, radar, lidar, thermal, and IMU
Lead improvements to multi-target tracking including track initiation, measurement-to-track association, and track state maintenance — with a focus on reliability in the hardest real-world conditions
Develop novel algorithms that incorporate noisy, incomplete measurements into a robust foundational understanding of the world
Build the tooling, metrics, and visualization infrastructure that lets the team move faster and evaluate progress rigorously
Set a high technical standard across the Perception team through design reviews, code reviews, and mentorship
Partner with teams across autonomy to ensure perception performance translates into system-wide safety and reliability
What You'll Bring
BS or MS with at least 8 years of industry experience, or PhD with at least 5 years, building and deploying state estimation and sensor fusion systems
Deep domain expertise across sensor fusion, state estimation, Kalman filtering, particle filtering, and multi-object tracking
Strong experience with deep learning and learned scene estimation — you know when to use classical methods and when learned approaches win
Exceptional C++ skills and a track record of writing highly maintainable, production-quality code in high-reliability systems
Qualification
Expertise in sensor fusionExceptional C++ skillsHistory of technical leadership
Required
BS or MS with at least 8 years of industry experience or PhD with at least 5 years
Expertise in sensor fusion, state estimation, Kalman filtering, particle filtering, and multi-object tracking
Strong experience with deep learning and learned scene estimation
Exceptional C++ skills
Ability to define metrics and design rigorous evaluations
History of technical leadership
Preferred
Behavior prediction and scene understanding at scale
Real-time perception systems under latency and compute constraints
Autonomous vehicle or robotics deployment in unstructured environments
Experience defining team-level technical roadmaps or leading multi-engineer efforts
