Research Intern: Learning Enhanced Goal and Intention Planning for Autonomous Driving

Research Intern: Learning Enhanced Goal and Intention Planning for Autonomous Driving - Honda Research Institute USA Job Number: P25INT-36 Honda Research Institute USA (HRI-US) is seeking a talented and motivated PhD-level research intern to explore learning-based decision-making for autonomous driving. The intern will develop models that predict high-level goals for ego planning and infer agent intentions from onboard sensory data, leveraging visual and semantic scene understanding. These predictions will inform behavior planning and interface with downstream planning modules to enable robust and adaptive decision-making in complex driving environments. Emphasis will be placed on developing real-time, deployable methods that can be seamlessly integrated into a full planning stack, with a focus on efficiency, reliability, and practical applicability. Mountain View, CA Key Responsibilities Conduct a survey of learning-based approaches for goal inference, intention learning, and behavior planning, including methods leveraging visual and semantic representations. Design and implement models that predict high-level driving goals and infer agent intentions from onboard sensor data (e.g., cameras). Develop representations that integrate visual and semantic information for structured and efficient decision-making. Optimize models for real-time performance, including considerations of latency, computational efficiency, and memory footprint. Interface predicted goals with downstream planning modules and evaluate their impact on end-to-end planning performance. Train and evaluate models using internal datasets and/or simulation environments. Benchmark and analyze trade-offs between model accuracy, inference speed, and system-level performance. Collaborate with HRI researchers to refine approaches and contribute to technical publications. Minimum Qualifications Ph.D. candidate in Robotics, Computer Science, Electrical Engineering, or a related field. Strong background in at least one of the following: machine learning, computer vision, decision-making, or autonomous driving. Experience with deep learning frameworks (e.g., PyTorch, TensorFlow). Experience working with vision-based models (camera-only perception or reasoning). Programming proficiency in Python (C++ is a plus). Ability to conduct independent research and prototype solutions. Bonus Qualifications Experience with goal prediction, intention learning, or behavior planning in autonomous systems. Familiarity with visual scene understanding and semantic reasoning. Experience working with large-scale driving datasets and offline evaluation pipelines. Experience with simulation environments for autonomous driving. Experience integrating learning-based models into planning or decision-making pipelines. Experience with foundation models (e.g., vision-language models) and fine-tuning large models for downstream tasks. Familiarity with adapting or fine-tuning large language models (LLMs) or multimodal models for structured reasoning or decision-making tasks. Prior research publications in robotics, machine learning, or computer vision. Years of Work Experience Required 0 Desired Start Date 9/14/2026 Internship Duration 3 Months #J-18808-Ljbffr Honda Research Institute USA