Research Scientist: E2E Autonomous Mobility
Honda Research Institute USA, Inc.
Summary
Honda Research Institute USA (HRI-US) is seeking a highly motivated Research Scientist to contribute end-to-end pipeline and learning technologies for autonomous mobility, spanning perception, scene understanding, planning and control. A successful candidate will have experience with one or more of the following: end-to-end driving model/stack, closed-loop performance improvement, end-to-end hardware experiments on real vehicles, vision-language-action models, world models, imitation or reinforcement learning for driving. The ideal candidate combines strong research capabilities with practical experience deploying models in real-world systems.
Job description
Key Responsibilities
- Research and develop end-to-end learning pipelines for autonomous mobility, spanning perception, scene understanding, and driving policy.
- Design and train end-to-end driving models, including vision-language-action models, foundation models, world models, and imitation or reinforcement learning-based policies.
- Improve closed-loop performance through simulation, data-driven evaluation, and iterative model refinement with simulation and hardware.
- Lead end-to-end hardware experiments on real vehicle platforms, from small-scale cars (e.g., RoboRacer; former F1TENTH) to full-scale vehicles.
- Contribute new research ideas and publish at top-tier venues (e.g., RAL, ICRA, IROS, CoRL, CVPR, ICCV, ECCV, NeurIPS).
Minimum Qualifications
- Ph.D. in Computer Science, Robotics, Electrical Engineering, or a related field
- 2+ years of hands-on experience in machine learning / deep learning with PyTorch or TensorFlow
- Research experience in one or more of: end-to-end driving model/stack, end-to-end training of differentiable modules, vision-language-action models, world models, imitation or reinforcement learning for driving, or closed-loop performance improvement
- Hands-on experience deploying and testing learned models on real vehicle hardware, including small-scale autonomous cars (e.g., RoboRacer; F1TENTH, MuSHR, Duckietown) and/or full-scale vehicles
- Strong programming skills in Python and/or C++
- Experience building end-to-end training and evaluation pipelines for perception or driving models
- Strong publication record in robotics, controls, computer vision, or machine learning
Bonus Qualifications
- Experience running closed-loop experiments on full-scale autonomous vehicles, including safety driver protocols and on-vehicle data logging
- Experience with ROS/ROS2, embedded compute (e.g., NVIDIA Jetson, DRIVE), and vehicle sensor stacks (camera, LiDAR, radar, IMU)
- Experience with model optimization for real-time on-vehicle inference (quantization, distillation, TensorRT)
- Experience with closed-loop simulation (e.g., CARLA, Waymax, NVIDIA Alpasim, Isaac sim) and sim-to-real transfer
- Experience with LLMs, VLMs, or foundation models for driving or scene understanding
- Experience with large-scale distributed training and multimodal datasets
Requirements
- Ph.D. in Computer Science, Robotics, Electrical Engineering, or a related field
- 2+ years of hands-on experience in machine learning / deep learning with PyTorch or TensorFlow
- Research experience in end-to-end driving model/stack, end-to-end training of differentiable modules, vision-language-action models, world models, imitation or reinforcement learning for driving, or closed-loop performance improvement
- Hands-on experience deploying and testing learned models on real vehicle hardware, including small-scale autonomous cars and/or full-scale vehicles
- Strong programming skills in Python and/or C++
- Experience building end-to-end training and evaluation pipelines for perception or driving models
- Strong publication record in robotics, controls, computer vision, or machine learning
Company overview
Honda Research Institute USA, Inc. (HRI-US) is North America's advanced research center established in 2003 as part of Honda's global research network. Based in San Jose, California with additional research labs in Columbus, Ohio, HRI-US serves as Honda's premier research facility focused on developing innovative solutions with direct applications to Honda's current and future technology roadmap. The institute employs a team of scientists and engineers who create technologies through scientific processes and apply them to real-world situations, addressing complex problems beyond abstract principles. HRI-US specializes in cutting-edge research areas including quantum technologies, nanomaterials, carbon capture technologies, energy and battery technologies, mobility-related human-computer interaction, software-defined mobility, and next-generation hydrogen fuel cells. The institute maintains an open innovation model, establishing strategic partnerships and alliances with academic institutions, public organizations, and private sector entities to foster innovation and drive technological advancement across multiple domains including computer science, material science, and human-machine interfaces.