Tech Lead, Perception Autonomy

Rivian Rivian · Auto · Palo Alto, CA · Autonomous Driving

Tech Lead for Perception Autonomy at Rivian, focusing on architecting and delivering production-grade ML pipelines for AV mapping, including multi-vehicle, lidar-free, and HD mapping/localization. Owns the end-to-end ML lifecycle from data acquisition to optimization and evaluation, with a strong emphasis on shipping scalable solutions.

What you'd actually do

  1. Own, architect, drive and deliver the overarching strategy to push the frontiers on mapping.
  2. Ship a production-grade, high-quality, robust, scalable mapping pipeline.
  3. Have a holistic understanding of the entire AV perception stack and work with the teams to define how we measure and monitor performance of the mapping system and its capabilities.
  4. Work with the team to drive progress in improving the performance.
  5. Establish rigorous evaluation and monitoring benchmarks.

Skills

Required

  • BS, MS, or PhD in Computer Science, Robotics, Electrical Engineering, or a highly related quantitative field.
  • 7+ years of professional experience building, scaling and shipping ML solutions
  • Proven track record of hands-on experience delivering a mapping system for Autonomous Vehicles at scale.
  • Experience with defining, driving and delivering a mapping strategy for AV.
  • Experience in shipping mapping systems with different sensor modalities in AV, especially both with and without lidar.
  • Holistic understanding of the entire AV perception stack.
  • Strong proficiency in Python and/or C++
  • Solid understanding of modern Perception pipelines, benchmarking tools, and infrastructure.
  • Demonstrated ability to drive progress across a complex system spanning multiple domains and components, across a distributed, cross-functional stack in a fast-paced environment.

Nice to have

  • Strong experience in AV mapping, especially from multiple vehicle passes and across time.
  • Strong experience in Lidar-free mapping for Autonomous Vehicles
  • Experience in building HD maps and SD maps.
  • Experience in Localization given a map, with and without lidar.
  • Experience in defining and driving a training data strategy, including defining data annotation guidelines, partnering effectively with in-house and external 3P annotation vendors.
  • Experience with multiple modalities (e.g., cameras, lidar, radar).
  • Experience with auto-labeling (lane, objects, etc)
  • Experience with onboard edge deployment, cloud inference architectures, and balancing compute/efficiency trade-offs.
  • Experience with optimization, hyper-scalability, high-performance inference engines like TensorRT.

What the JD emphasized

  • shipping production-grade pipelines
  • push the frontiers on mapping
  • multi-vehicle mapping (with and without lidar)
  • 3D/4D mapping without lidar
  • HD maps & localization
  • end-to-end ML lifecycle & data flywheel
  • metrics definition
  • evaluation
  • model performance optimization
  • feedback loop
  • AV mapping at scale
  • delivering a mapping system for Autonomous Vehicles at scale
  • defining, driving and delivering a mapping strategy for AV
  • shipping mapping systems with different sensor modalities in AV, especially both with and without lidar
  • holistic understanding of the entire AV perception stack
  • benchmarking tools
  • infrastructure
  • drive progress across a complex system spanning multiple domains and components, across a distributed, cross-functional stack in a fast-paced environment
  • Lidar-free mapping for Autonomous Vehicles
  • building HD maps and SD maps
  • Localization given a map, with and without lidar
  • defining and driving a training data strategy
  • auto-labeling
  • onboard edge deployment
  • cloud inference architectures
  • balancing compute/efficiency trade-offs
  • optimization
  • hyper-scalability
  • high-performance inference engines like TensorRT

Other signals

  • shipping production-grade ML pipelines
  • end-to-end ML lifecycle & data flywheel
  • multi-vehicle mapping
  • 3D/4D mapping without lidar
  • HD maps & localization