Staff Software Engineer, Machine Learning (Perception)

Who We Are

Aurora (Nasdaq: AUR) is delivering the benefits of self-driving technology safely, quickly, and broadly to make transportation safer, increasingly accessible, and more reliable and efficient than ever before. The Aurora Driver is a self-driving system designed to operate multiple vehicle types, from freight-hauling semi-trucks to ride-hailing passenger vehicles, and underpins Aurora Horizon and Aurora Connect, its driver-as-a-service products for trucking and ride-hailing. Aurora is working with industry leaders across the transportation ecosystem, including Toyota, FedEx, Volvo Trucks, PACCAR, Uber, Uber Freight, U.S. Xpress, Werner, Covenant, Schneider, and Ryder. For Aurora’s latest news, visit aurora.tech and @aurora_inno on Twitter.

Aurora hires people who are excited to build the future of transportation. We are looking for machine learning and computer vision experts to develop and implement the core algorithms for interpreting sensor data that enables intelligent driving decisions. You will be developing and applying learning-based algorithms related to sequential prediction, recursive state estimation, and classification for object detection and tracking.

In this role, you will:

• Have an opportunity to tackle end-to-end Perception challenges across various areas relevant to the autonomy system. This includes working on detection and tracking of traffic actors, their action recognition, and semantic understanding of various traffic scenes that the Aurora Driver encounters
• Research and develop state-of-the art deep learning/machine learning models to improve our perception solutions under challenging and diverse scenarios (e.g., on highways and city streets, long-range and multi-sensor object detection, detection under degraded sensors, etc.)
• Collaborate with other members to ideate and improve the Perception system from problems encountered in on-road and simulated testing, and produce robust, well-tested software
• Guide the software development from prototype to production, including training and analysis on large-scale data sets, as well as deploying production-ready solutions on a real-time AV platform that powers the Aurora Driver
• Example projects that a candidate could potentially contribute to (list not exhaustive):
• Design and implementation of state-of-the-art object detection from multimodal sensors
• Design algorithms to extract, leverage, and fuse complementary information from different sensors
• Design and implement state-of-the-art ML systems for object tracking leveraging expertise in recursive state estimation, temporal models, structured prediction, and sequential models
• Analysis of failure modes and implementation/testing of solutions to address the surfaced issues
• Build and scale ML pipelines for quick experimentation, as well as large-scale training and testing

Required Qualifications

• Expertise in one or more relevant focus areas of ML: deep learning, computer vision, recursive state estimation, structured prediction, optimization, and others
• Strong publication record at top-tier robotics/computer vision conferences/journals, or significant industry experience in relevant fields (robotics, computer vision, self-driving technology)
• Strong grasp of fundamentals: linear algebra, discrete and continuous optimization, supervised and unsupervised methods, generative and discriminative methods
• Good knowledge of C++ and Python
• Experience with CUDA, cuDNN, TensorRT, and other related frameworks
• Prior work on optimizing and speeding up deep networks

Desirable Qualifications 

• Experience with PyTorch, Tensorflow, and/or other deep learning frameworks
• Advanced knowledge of C++ a big plus
• Experience with applying CV/ML to problems in robotics
• Prior work on deployment of CV/ML models on large fleets
• Prior experience in organizations where strong cross-team collaboration was needed
• Strong presentation and communication skills

The base salary range for this position is $189K-$302K per year.  Aurora’s pay ranges are determined by role, level, and location. Within the range, the successful candidate’s starting base pay will be determined based on factors including job-related skills, experience, qualifications, relevant education or training, and market conditions. These ranges may be modified in the future. The successful candidate will also be eligible for an annual bonu