ML Software Engineer, Behavior Planning

As a Machine Learning Software Engineer at Aurora, you'll primarily focus on developing machine learning algorithms for motion planning, ensuring that our self-driving vehicles generate safe and comfortable trajectories. You'll also work on policies for multi-actor interactions and implement robust systems that predict and adapt to the behavior of other road users.

To qualify for the Machine Learning Software Engineer role at Aurora, you'll need a BS, MS, or PhD in Robotics, Machine Learning, Applied Mathematics, or Computer Science, along with at least a year of experience in a production environment. Strong programming skills in C++ or Python and familiarity with modern machine learning frameworks will also be essential.

Beyond the basic qualifications, beneficial skills for a Machine Learning Software Engineer at Aurora include experience with reinforcement learning, familiarity with GPU toolkits like TensorFlow or PyTorch, and a strong grasp of both classical and modern ML approaches, such as transformer networks and convolutional neural networks.

The salary range for the Machine Learning Software Engineer position at Aurora is $105K to $157K per year, with final offers depending on factors like experience, skills, and market conditions. In addition to base pay, you’ll be eligible for an annual bonus and equity compensation.

Absolutely! Aurora is committed to fostering diversity and inclusion within its workforce. We welcome applications from candidates of all backgrounds and consider various factors, including race, gender identity, and disability status, ensuring a fair and accessible hiring process.

In your response, highlight specific frameworks like TensorFlow, PyTorch, or JAX that you have experience with. Discuss the types of projects you've worked on, focusing on how you've applied these tools to solve complex problems in robotics or motion planning.

Explain your methodology, emphasizing your understanding of trajectory generation and the importance of safety in interactions with other road users. Share examples of algorithms you’ve developed or studied, discussing how they enhance vehicle performance in dynamic environments.

Discuss the principles of reinforcement learning and connect them to decision-making in robotics. Provide examples of how you’d implement RL algorithms to improve a vehicle’s ability to make real-time decisions based on environmental feedback.

Be sure to tell a compelling story about a specific technical challenge, detailing the problem, your analytical process, the solution you implemented, and the positive outcome. Highlight your problem-solving and critical-thinking skills.

Mention any relevant conferences, journals, online courses, or communities you follow. Emphasize your proactive approach to continuous learning and how you integrate new knowledge into your work.

Discuss your systematic approach to debugging, perhaps by breaking down the algorithm into smaller parts, using logging, or employing visualization techniques to understand how data flows through your model.

Explain your first steps in diagnosing the failure, including data analysis and reassessment of the algorithm's logic. Discuss your iterative approach to refining the algorithm while maintaining a robust safety framework.

List languages like C++ or Python and describe concrete examples where you utilized these languages effectively in projects related to machine learning or robotics, focusing on your contributions and the impact of your work.

Share an example that highlights teamwork, emphasizing your role, communication skills, and how you contributed to the project's success. Focus on collaborative problem-solving and knowledge sharing.

Articulate your passion for autonomous vehicles, your alignment with Aurora's mission, and how your skills and aspirations fit within the company's innovative environment. Make it personal and specific to Aurora.