Machine Learning Software Engineer, Motion Planning: Behavior

$130,000 - $195,000/Yr

Aurora Innovation - Pittsburgh, PA

Full-time - Mid Level

Pittsburgh, PA

Professional, Scientific, and Technical Services

About the position

As a Machine Learning Software Engineer on the Motion Planning Team at Aurora, you will play a crucial role in developing advanced algorithms that enable self-driving vehicles to navigate safely and efficiently. Your primary focus will be on building machine learning approaches for motion planning, which involves creating, generating, and selecting trajectories that ensure both comfort and safety for passengers and other road users. You will be tasked with developing policies and plans that effectively manage interactions between multiple actors in dynamic environments, particularly under conditions of uncertainty. This role requires a deep understanding of how to interleave motion planning with the forecasting of other agents, ensuring that our self-driving systems can anticipate and react to the actions of other vehicles and pedestrians in real-time. In addition to algorithm development, you will implement production-grade solutions that are integral to the functionality of our self-driving technology. This position offers the opportunity to work at the forefront of autonomous vehicle technology, collaborating with industry leaders and contributing to a safer and more efficient transportation ecosystem. You will be part of a diverse team that values innovative thinking and is committed to making mobility accessible for all. Your contributions will directly impact the development of our driver-as-a-service products, which are designed to operate across various vehicle types, from freight-hauling trucks to passenger vehicles.

Responsibilities

• Build machine learning approaches for motion planning to create, generate, and select comfortable and safe trajectories for self-driving vehicles.
• Develop policies and plans to manage multi-actor interactions and plans under uncertainty.
• Develop approaches that interleave motion planning with forecasting of other agents.
• Develop and implement production-grade algorithms for the Driver.

Requirements

• BS, MS, or PhD in Robotics, Machine Learning, Applied Mathematics, Computer Science or related field.
• 1+ years of industry experience building software in a production environment.
• Strong programming and software design skills ideally in C++ and/or Python.
• Knowledge of classical and modern ML approaches such as transformer networks, linear models, recurrent models, convolutional neural networks.
• Familiarity with modern ML and GPU toolkits such as PyTorch, TensorFlow, JAX, TensorRT, and/or CUDA.

Nice-to-haves

• Experience with imitation and reinforcement learning and how machine learning can be used in robotics applications.
• Experience in designing, building, and productionizing robotics systems, especially decision making systems, motion planning algorithms, state estimation algorithms, probabilistic modeling, or statistical techniques.

Benefits

• Annual bonus
• Equity compensation
• Comprehensive health benefits