Auto-labelling is a foundational pillar of the Autonomy stack. In this Senior ML Engineer role, you will play a key role in delivering high-quality, scalable auto-labeling models. This includes training, optimizing and shipping auto-labeling models in the Autonomy stack. Use cases include mapping, lanes auto-labelling, object auto-labelling as well as other critical applications. You will ship production-grade models that push the boundaries of what’s possible. As such, you will also contribute to the whole end-to-end ML lifecycle & data flywheel of this effort: data acquisition, metrics definition, evaluation, model performance optimization, feedback loop. A key part of the role is especially dedicated to lidar-free auto-labeling, i.e. ship auto-labeling models that do not require lidar data.

• Education: BS, MS, or PhD in Computer Science, Robotics, Electrical Engineering, or a highly related quantitative field.
• Experience: 5+ years of professional experience building and scaling ML solutions, with a strong focus on the following:
  • AV auto-labeling system at scale: Proven track record of hands-on experience delivering auto-labeling models for Autonomous Vehicles at scale. Auto labeling for mapping, lanes auto-labelling and/or object auto labelling.
  • Perception stack: solid understanding of the AV perception stack.
  • System engineering: Strong proficiency in Python alongside a solid understanding of modern Perception pipelines, benchmarking tools, and infrastructure.
  • Execution: Demonstrated ability to drive progress across a complex, multi-domain system, in a fast-paced environment.

• Experience in one of the following auto-labeling applications: mapping, lanes auto-labelling or object auto-labelling.
• Experience in Lidar-free auto-labeling
• Experience in mapping, especially from multiple vehicle passes and/or lidar-free mapping.
• Experience in complex,multi-modal, large-scale data flywheel
• Experience with multiple modalities (e.g., cameras, LiDAR, Radar).
• Experience with onboard edge deployment, cloud inference architectures, and balancing compute/efficiency trade-offs