About the job
Building robust ML and/or mapping systems that work with real data (cameras, LiDAR, etc.) in uncertain environments, and a strong desire to contribute to the future of robot navigation for logistics and transportation.
Responsibilities
Research, develop, and implement state-of-the-art online mapping models and algorithms.
Analyze and characterize the performance of the online mapping system, identifying opportunities for architecture, data or evaluation improvements in a E2E ML system.
Work cross functionally with other ML teams to integrate our models into centralized architectures.
Collaborate with stakeholders across autonomy, infrastructure, and systems teams on online mapping needs and requirements.
Qualifications
Minimum
Proven record of solving in-production ML problems and making tradeoffs between data, model and evaluation.
Deep understanding of ML fundamentals with hands-on experience in training and evaluating modern ML models with applications in AV, robotics, mapping, computer vision, or related areas.
Prioritize impact and practicality, and make decisions to ensure on-time delivery of solutions.
Able to quickly iterate and experiment to cut through the noise and make well-informed, data-driven decisions.
Experience with robotics-related ML applications, 3D geometry.
Strong Python skills with experience in deep learning frameworks, e.g., PyTorch, TensorFlow, or Jax.
Preferred
Proficiency in working with complex multi-component systems.
Experience in building ML pipelines and optimizing/productizing ML models.
Familiarity with modern ML tools and infrastructure such as distributed training and ML compilers.
Demonstrated research publications in top conferences (e.g. NeurIPS, ICLR, ICML, CVPR, RSS, CoRL, ICRA).
Deep understanding of 3D geometry and state estimation fundamentals.