Senior/staff Machine Learning Engineer, Sensor Simulation

Nuro Nuro · Robotics · Mountain View, CA · Sensor Data and Calibration

Nuro is seeking a Senior/Staff Machine Learning Engineer to join their Sensor Data and Calibration team. The role involves researching, developing, and implementing state-of-the-art synthetic sensor simulation methods and algorithms, with a focus on generating realistic synthetic sensor data (images, lidar, radar) using ML techniques like NeRF or Gaussian splatting. The engineer will analyze the realism and utility of this data and collaborate with other teams to address critical questions about sensor data and autonomy performance.

What you'd actually do

  1. Research, develop, and implement state-of-the-art synthetic sensor simulation methods
  2. Analyze and characterize the realism and utility of synthetic sensor data
  3. Answer critical questions about sensor data and autonomy performance
  4. Collaborate with stakeholders across autonomy, infrastructure, and systems teams on map needs and requirements

Skills

Required

  • Python
  • PyTorch
  • TensorFlow
  • Jax

Nice to have

  • Deep understanding of 3D geometry and state estimation fundamentals
  • Proficiency in systems coding
  • Experience in simulating/modeling real sensors (camera, lidar, radar, IMU, etc), including noise modeling
  • Experience in modern ML graphics techniques, e.g., NeRF, Gaussian Splatting, and/or generative models.
  • Experience in building ML pipelines and optimizing/productizing ML models
  • Demonstrated research publications in top conferences (e.g. NeurIPS, ICLR, ICML, CVPR, RSS, CoRL, ICRA)

What the JD emphasized

  • hands-on experience in the research, development, and implementation of machine learning methods
  • Deep understanding of ML fundamentals with hands-on experience in training and evaluating modern ML models
  • Experience in modern ML graphics techniques, e.g., NeRF, Gaussian Splatting, and/or generative models.

Other signals

  • synthetic sensor simulation models
  • machine learning methods for generating synthetic sensor data
  • photorealistic images, realistic lidar and/or radar