Perception Engineer

Motional•Boston, MA

8h•Hybrid

About The Position

Develop computer vision and machine learning solutions for autonomous vehicle perception systems, including recognition of objects and signals such as emergency vehicles, school buses, authority figures, and traffic-light states. Design, develop, implement, and debug production perception software using Python and C++, and integrate perception algorithms and trained machine learning models with on-car software. Architect workflows for preparing data from multiple sensors, such as cameras and LiDAR, and training machine learning models. Design evaluation processes and custom testing tools to measure prediction consistency, processing speed, and overall system safety. The role is based at Motional’s Boston headquarters, with telecommuting permitted. Motional is a driverless technology company working to make autonomous vehicles safe, reliable, and accessible. The company describes its mission as people-first and focused on safer roadways, equitable transportation, and improving communities. It is working toward commercialization and is a joint venture between Hyundai Motor Group and Aptiv, with operations in the U.S. and Asia.

Requirements

Bachelor’s degree in Computer Science, Electrical Engineering, or a closely related technical field.
Two years of industry experience in automotive industry software development.
Experience in production-level software development.
Experience with model deployment, integration, and validation for robotics.
Experience testing and debugging algorithms.
Experience with machine learning applications.
Experience with Python and C++.
Experience with PyTorch.
Experience with GPU training, acceleration, and inference optimization.
Experience with TensorRT/CUDA.
Experience with CI/CD pipelines and Git.

Responsibilities

Develop computer vision and machine learning solutions for autonomous vehicle perception, including object and signal recognition.
Design, develop, implement, and debug features in production perception software using Python and C++.
Integrate perception algorithms and trained machine learning models with on-car software.
Architect workflows for preparing data from multiple sensors, such as cameras and LiDAR, and training machine learning models.
Design evaluation processes and custom testing tools to measure prediction consistency, processing speed, and overall system safety.
Drive project-level impact through technology solutions.
Formulate problems, architect solutions, and design processes.
Train and deploy new AI models.
Analyze model performance and identify improvements.
Research and evaluate emerging technologies to enhance AI models.
Prototype, evaluate, implement, and iterate on solutions.
Design, develop, implement, and fix features in the codebase.