Senior / Staff ML Onboard Optimization Engineer

Waabi Waabi · Robotics · US & Canada, Dallas, TX +4 · Remote · Autonomy & Algorithms

This role focuses on optimizing and deploying machine learning models for onboard compute systems in autonomous vehicles. It involves expanding the deployment pipeline, optimizing models using frameworks like TensorRT, and creating custom CUDA kernels for inference. The engineer will also profile model runtime and memory to identify performance bottlenecks.

What you'd actually do

  1. Expand the model deployment pipeline to new GPUs and embedded systems for the next generation of our onboard compute system.
  2. Use frameworks such as TensorRT and modelopt to optimize the models running on the truck.
  3. Create and benchmark new CUDA kernels for inference.
  4. Comprehensively profile model runtime and memory to pinpoint performance bottlenecks.

Skills

Required

  • Python
  • C++
  • Rust
  • PyTorch
  • PyTorch Profiler
  • NVIDIA Nsight
  • Nvidia embedded platforms (Jetson or Thor)

Nice to have

  • model compilation and exporting
  • TensorRT
  • custom CUDA kernels
  • Bazel build systems

What the JD emphasized

  • next generation of our onboard compute system
  • CUDA kernels for inference

Other signals

  • optimizing models for onboard compute
  • expanding model deployment pipeline
  • creating and benchmarking CUDA kernels for inference
  • profiling model runtime and memory