Machine Learning Engineer - Perception Attributes
hace 21 días
Boston
Job DescriptionThe Perception team at Zoox is fundamental to our autonomous vehicle technology, creating the understanding of the world for our self-driving robots. We enable safe and efficient navigation in complex environments through sophisticated detection, classification, and tracking systems. As a Machine Learning Engineer on the Attributes team within the Perception department, you will take ownership of developing and enhancing sophisticated behavioral models for various road users, including vehicles, pedestrians, and cyclists. Your work will focus on creating and maintaining robust perception attribute models that generate critical signals for our autonomous driving stack. These signals are essential inputs that enable our Prediction and Planning teams to make intelligent, safe driving decisions for our autonomous vehicles. In this role, you will bridge the crucial gap between raw perception data and autonomous decision-making, working closely with cross-functional teams to optimize model outputs and enhance our system's overall perception capabilities. Your contributions will directly impact the safety and effectiveness of Zoox's autonomous vehicle platform, advancing our mission of revolutionizing urban mobility. Your work in the Attributes team is fundamental to Zoox's autonomous driving capabilities. The models you develop and the impact you will have will serve as the foundation for how our vehicles understand and interact with other road users, directly contributing to the safety and effectiveness of our autonomous driving system. By creating reliable and accurate behavioral models, you enable Zoox's vehicles to make smart, safe decisions in complex urban environments, bringing us closer to our goal of revolutionizing urban mobilityIn this role, you will: • Work with Data Labeling and Ontology teams on data labeling and ontology definitions of the road users in different attributes and generate auto-labeling or data mining strategies for different attributes., • Lead the development of sophisticated behavioral models for vehicles, pedestrians, and cyclists as a key member of the Attributes team within Zoox's Perception department., • Create and maintain Perception attribute models that generate essential signals, enabling our autonomous vehicles to understand and predict the behavior of various road users., • MS/PhD in Computer Science or related fields with a minimum of 5 years of relevant experience, • Experience with training and deploying Deep Learning models, • Experience with production Machine Learning pipelines: dataset creation, training frameworks, metrics pipelines, • Fluency programming in Python and extensive experience with algorithm design, • Strong mathematics skillsBonus Qualifications, • Familiarity of VLMs/VLAs/ViTs, • Experience with large model distillation in a production environment