Research Engineer, Learnable Planner (integration)

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

Research Engineer role focused on integrating ML models into the production planning stack for autonomous trucks, developing simulation pipelines, and improving planner architecture. Requires MS/PhD in a related field and experience with ML-based planning techniques, deep learning frameworks, and programming languages like Python, Rust, C++, or CUDA. Bonus for production deployment experience and model iteration.

What you'd actually do

  1. Integrate cutting-edge ML models in production planning stack from development to validation, deployment, and monitoring
  2. Develop necessary interfaces and pipelines in simulation for testing prototype or production planning models
  3. Work closely with motion planning sub-teams and research scientists to improve our planner architecture and develop rich and novel representations that can facilitate end-to-end solutions
  4. Champion engineering excellence, ensuring high-quality, well structured and tested code.
  5. Stay up-to-date with the latest advancements in the field of artificial intelligence, machine learning, computer vision, and self-driving technologies, and apply insights from the literature.

Skills

Required

  • MS/PhD in machine learning, computer science, engineering, or a related field
  • Experience in ML-based or classical techniques for planning/decision making (e.g., imitation and reinforcement learning, optimization-based approaches, search methods, probabilistic reasoning)
  • Solid understanding of computing fundamentals, including code efficiency
  • Experience in deep learning frameworks such as PyTorch
  • Proficiency in Python, Rust, C++ and/or CUDA

Nice to have

  • Experience deploying ML/DL models to a production motion planning or related robotics stack
  • Experience in iterating on a model including evaluation, introspection and fine-tuning
  • Strong grasp of machine learning literature, including current trends and state-of-the-art techniques
  • Comfortable with model compilation and exporting, lower level concepts like TensorRT, CUDA kernels

What the JD emphasized

  • integration of new AI technologies into our autonomy/planner stack
  • fully driverless autonomous trucks
  • single AI system that learns end-to-end and in a provably safe manner
  • high-fidelity, closed-loop simulator, Waabi World
  • ML-based or classical techniques for planning/decision making
  • taking research ideas and turning them into practical solutions for real-world applications
  • deploying ML/DL models to a production motion planning or related robotics stack
  • iterating on a model including evaluation, introspection and fine-tuning

Other signals

  • integrating ML models into production planning stack
  • developing interfaces and pipelines in simulation for testing prototype or production planning models
  • improving planner architecture and developing rich and novel representations that can facilitate end-to-end solutions
  • applying insights from the literature
  • working with large datasets from various sources as well as Waabi World, our high-fidelity simulator