Principal Autonomous Vehicle Engineer, Perception

The challenge

The truck runs revenue freight across the operational corridors. New corridors and new weather conditions surface perception regressions weekly. The job is keeping fleet-wide metrics stable while the pipeline keeps improving. The gap between current performance and the operational target shows up across dozens of edge case categories — construction zones, emergency vehicles, occlusion at distance, low-sun glare, tunnel ingress and egress, rain backscatter on LiDAR, novel objects, and sensor degradation across weather conditions.

Role overview

Raw sensor data in, tracked object output out -- the entire perception pipeline is your technical domain. LiDAR point clouds, camera images, and radar returns get fused into the 3D world model that the planner consumes. Edge case robustness is treated as a category of work, addressed through hard-example mining, evaluation pipeline improvements, and targeted data collection rather than scenario-by-scenario fixes.

Specializations

Technical requirements

• MS or PhD in Computer Science, Robotics, or Electrical Engineering
• 10+ years in perception systems for autonomous vehicles or robots
• Published work or production experience with multi-sensor fusion architectures
• Deep experience with point cloud processing (PCL, Open3D, or custom pipelines)
• Proficiency in C++ and Python for real-time perception code
• Track record of improving perception metrics on real-world driving data

Key responsibilities

• Own the perception pipeline architecture for a Class 8 autonomous trucking platform
• Develop and deploy 3D object detection, semantic segmentation, and multi-object tracking models for real-time inference at highway speeds
• Develop sensor fusion algorithms that combine LiDAR, camera, and radar data into coherent tracked objects
• End-to-end delivery across training infrastructure, model optimization, on-vehicle deployment, and production monitoring
• Drive edge case robustness across construction zones, adverse weather (rain backscatter, snow occlusion, low-sun glare), occlusion at distance, and emergency vehicle response through hard-example mining, targeted data collection, and evaluation pipeline improvements
• Define labeling taxonomies and ground truth quality standards that feed the automated hard-example mining pipeline
• Collaborate with the planning team on perception output format and latency requirements

Compensation & Benefits

• Base salary: $270,000 - $310,000
• Meaningful equity stake (pre-IPO company, 4-year vest)
• Annual bonus target 15%
• Relocation package for Bay Area candidates
• Paid parental leave (16 weeks)