Software Engineer, Perception

What You'll Do

• Applying machine learning to solve challenging perception problems for autonomous systems (e.g. object detection, semantic segmentation, instance segmentation, dense depth, optical flow, tracking, etc.).
• Driving the architecture, deployment, and performance characterization of our deep learning models.
• Refining and optimizing models for low-latency inference on embedded hardware.
• Designing and building cloud-based training and labeling pipelines.
• Collaborating with the hardware and embedded teams on sensor selection and vehicle packaging given safety requirements.
• Writing performant, well-tested software, and improving code quality of the entire Autonomy team through code and design reviews.

What You'll Bring

• 5+ years of experience in software development for problems involving computer vision, machine learning, and robotic perception techniques.
• Foundational understanding of deep learning: model layer design, loss function intuition, training best practices.
• Experience handling large datasets efficiently and organizing them for training and evaluation.
• Experience curating synthetic and real-world image datasets for training.
• Strong proficiency in modern C++ and Python and experience writing efficient algorithms for resource-constrained embedded systems.
• Ability to thrive in a fast-moving, collaborative, small team environment with lots of ownership.
• Excellent analytical, communication, and documentation skills with demonstrated ability to collaborate with interdisciplinary stakeholders outside of Autonomy.
• An eagerness to get your hands dirty by testing your code on real robots at real customer farms (gives “field testing” a whole new meaning!).

What Makes You a Strong Fit

• Experience architecting multi-sensor ML systems from scratch.
• Experience with compute-constrained pipelines: optimizing models to balance the accuracy vs. performance tradeoff, leveraging TensorRT, model quantization, etc.
• Experience implementing custom operations in CUDA.
• MS or PhD in Robotics, Computer Science, Computer Engineering, or a related field.
• Publications at top-tier perception/robotics conferences (e.g. CVPR, ICRA, etc.).
• Passion for sustainable agriculture and electric vehicles.