Research Intern: Camera-Based Perception for Marine Robotics

Honda Research Institute USA; San Jose (CA); On-site; USD 34,000 - 48,000; Part time

Job summary

Honda Research Institute USA in San Jose, CA is seeking a self-motivated research intern to advance camera-based perception for marine robotics. You will explore areas such as open-world perception, 3D reconstruction, and sim-to-real transfer using foundation models, with hands-on work on on-water datasets and ROS 2 pipelines.

Ideal candidates are M.S. students in robotics/CS/EE, have DL experience in CV, and strong Python or C++ skills; prior work with real camera data and publications is a

Qualifications

• M.S. candidate in robotics, CS, EE, or a related field.
• Experience training deep learning models for computer vision tasks (segmentation, detection, depth).
• Strong programming skills in Python or C++ and hands-on data handling.

Responsibilities

• Train and evaluate perception models on marine datasets and reference sensors.
• Integrate models into a ROS 2 pipeline for on-vessel testing.
• Document results and contribute to publications.
• Collaborate with researchers on foundation model research and sim-to-real transfer.

Skills

Python

C++

Deep learning

Computer vision

Education

M.S. candidate in robotics, CS, EE, or related field

Job description

Research Intern: Camera-Based Perception for Marine Robotics - Honda Research Institute USA

Research Intern: Camera-Based Perception for Marine Robotics

Job Number: P26INT-30

Honda Research Institute USA (HRI-US) is seeking a self-motivated research intern to join our Intelligence Marine Robotics team in advancing camera-based perception for marine robots. Depending on your interests and project needs, you will focus on one or more research areas: marine foundation models for open-world scene understanding (object detection, segmentation, and depth estimation), 3D reconstruction and mapping, or sim-to-real transfer using generative models.

San Jose, CA

Key Responsibilities

• Use on-water datasets collected with our full-scale test vessel for training and evaluation.
• Evaluate perception accuracy against reference sensors such as radar and GNSS/INS.
• Integrate models into a ROS 2 pipeline for on-vessel testing.
• Document results and contribute to publications.

Minimum Qualifications

• M.S. candidate in robotics, computer science, electrical engineering, or a related field.
• Experience training deep learning models for computer vision tasks such as segmentation, detection, or depth estimation.
• Strong programming skills in Python or C++.
• Hands-on experience working with real-world camera data.

Bonus Qualifications

• Ph.D. candidate in a related field.
• Experience with self-supervised vision foundation models or transformer-based segmentation.
• Experience with open-vocabulary or open-set detection/segmentation, or vision-language models.
• Experience with monocular depth estimation or panoptic segmentation.
• Experience with 3D reconstruction or mapping.
• Experience with visual-inertial SLAM.
• Experience with diffusion models or other generative models, or with sim-to-real transfer.
• Experience with camera-radar or camera-LiDAR sensor fusion.
• Experience with ROS 2 and deploying models on robotic platforms.
• Experience working with GNSS and IMU data.
• Experience with marine robotics or outdoor field robotics data.
• Publication record in computer vision or robotics venues (e.g., CVPR, ICCV, ICRA, IROS).

Years of Work Experience Required 0

Desired Start Date May 2027

Internship Duration 3 Months

Position Keywords Perception; Computer Vision; Marine Robotics; Foundation Models; Open-World Perception; Depth Estimation