Geometric Map Generation Engineer (3D SLAM)

Aurora; San Francisco (CA); Hybrid; USD 139,000 - 223,000; Full time

Benefits offered by this job

Annual bonus

Equity

Benefits

Job summary

Aurora is hiring a Software Engineer to build the 3D spatial foundation powering the Aurora Driver. You will own algorithms processing millions of miles of lidar logs into automated offline maps, tackling speed and scaling challenges across new routes.

You will work on geometric map-generation algorithms, optimize pose-graph methods, and improve distributed map-build performance, collaborating across teams to define data contracts and integration requirements.

Qualifications

• BS/MS/PhD in CS/Robotics/EE or equivalent practical experience.
• 3+ years of professional software development experience in C++.
• Hands-on experience with 3D geometry and point-cloud processing (registration, pose-graph, nonlinear optimization).
• Strong grasp of linear algebra, spatial transforms, numerical optimization and probability.
• Experience debugging and improving data quality across large datasets or pipelines.

Responsibilities

• Design, implement and own geometric map-generation algorithms, including sensor optimization and large-scale pose-graph optimization.
• Identify and build a system that is robust to failure modes and ship algorithmic fixes with measurable success criteria.
• Improve runtime and scalability of large-scale distributed map-build pipeline stages.
• Collaborate with neighboring team on defining data contracts and integration requirements.

Skills

C++

3D geometry

Linear algebra

Data quality

Debugging

Education

BS/MS/PhD in CS/Robotics/EE

Tools

ICP

Scan matching

Pose-graph optimization

Lidar-based SLAM

Job description

Aurora is hiring a Software Engineer to build the 3D spatial foundation powering the Aurora Driver. You will own algorithms processing millions of miles of lidar logs into automated offline maps, tackling speed and scaling challenges across new routes.

You will work on geometric map-generation algorithms, optimize pose-graph methods, and improve distributed map-build performance, collaborating across teams to define data contracts and integration requirements.