Research Intern: Multimodal Learning for Human-AI Team State Sensing

Honda Research Institute USA; San Jose (CA); On-site; USD 34,000 - 55,000; Full time

Job summary

Honda Research Institute USA invites a research intern to develop multimodal fusion methods for understanding human and AI behavior in group interactions. You will explore integrating language, speech, visual signals, and interaction context through learning representations across modalities and participants.

candidates should be PhD students in CS/Robotics/Cognitive Science with a strong publication record and Python/Machine Learning framework experience.

Qualifications

• Current Ph.D. student in Computer Science, Robotics, Cognitive Science or a related field.
• Familiarity with multimodal fusion, representation learning, cross-modal alignment, foundation models, or social signal processing.
• Proficiency in Python and experience with established machine learning frameworks.
• Proven publication record in machine learning, NLP, CV, or HCI venues.

Responsibilities

• Develop multimodal fusion methods for integrating language, speech, and visual signals.
• Learn representations that capture information across modalities, participants, and time.
• Evaluate fusion approaches for conflicting or missing signals.
• Examine model generalization, interpretability, and robustness across interaction settings.
• Present research progress and contribute to publications, prototypes, or demonstrations.

Skills

multimodal fusion

representation learning

cross-modal alignment

foundation models

social signal processing

Python

Education

Ph.D. student in CS/Robotics/Cognitive Science

Job description

Research Intern: Multimodal Learning for Human-AI Team State Sensing - Honda Research Institute USA

Research Intern: Multimodal Learning for Human-AI Team State Sensing

Job Number: P26INT-18

Honda Research Institute USA (HRI-US) is seeking a research intern to develop multimodal fusion methods for understanding human and AI behavior in group interactions. The intern will investigate how information from language, speech, visual behavior, and interaction context can be effectively integrated. They will explore machine learning approaches for combining complementary and sometimes incomplete signals across modalities and participants. The intern may also borrow concepts from psychology, cognitive science, or organizational interaction research to inform the pipeline design and evaluation of multimodal models.

San Jose, CA

Key Responsibilities

• Develop multimodal fusion methods for integrating language, speech, and visual signals.
• Learn representations that capture information across modalities, participants, and time.
• Evaluate fusion approaches for complementary, missing, uncertain, or conflicting signals.
• Examine model generalization, interpretability, and robustness across interaction settings.
• Present research progress and contribute to publications, prototypes, or research demonstrations.

Minimum Qualifications

• Current Ph.D. student in Computer Science, Robotics, Cognitive Science or a related field.
• Familiarity with multimodal fusion, representation learning, cross-modal alignment, foundation models, or social signal processing.
• Proficiency in Python and experience with established machine learning frameworks.
• Proven publication record in machine learning, natural language processing, computer vision, or human-computer interaction venues.

Bonus Qualifications

• Experience in social signal processing and affective computing with publications in top-tier conferences.
• Research experience with group interaction or human-AI teaming.
• Familiarity with fine-tuning vision-language models using human data.
• Experience evaluating models across datasets, tasks, participants, and interaction contexts.
• Experience in interpretable modeling or incorporating concepts from psychology or cognitive science.

Years of Work Experience Required

0

Desired Start Date

May 2027

Internship Duration

3 Months

Position Keywords

Multimodal Machine Learning, Human-AI Teaming, Computer Vision, Psychology