Research Scientist: AI for Collective State Sensing ...

Honda Research Institute USA, Inc.; San Jose (CA); On-site; USD 100,000 - 130,000; Full time

Job summary

Honda Research Institute USA, Inc. is looking for a Research Scientist in San Jose, CA to develop AI methods for understanding human interactions through multimodal data. The ideal candidate will have significant experience in AI/ML, human behavior modeling, and translating research into prototypes. This role offers the chance to work in interdisciplinary teams and contribute to cutting-edge AI systems.

Candidates should hold a Ph.D. and have a strong publication record in relevant fields.

Qualifications

• Ph.D. in computer science, AI or a closely related field.
• Strong research record in reputable conferences and journals in AI/ML.
• Experience developing and applying deep learning models.

Responsibilities

• Develop AI models for collective state sensing and group-level behavior modeling.
• Create methods for learning from noisy human behavioral signals.
• Model social roles and coordination in human-AI systems.

Skills

Multimodal learning

Deep learning models

Computer vision

Natural language processing

Signal processing

Education

Ph.D. in computer science, AI or related field

Tools

Python

PyTorch

TensorFlow

Job description

Job Number: P25F07

Honda Research Institute USA (HRI-US) is seeking a Research Scientist to develop AI methods for sensing, modeling, and interpreting collective states and behaviors in human‑human, human‑AI, and AI‑mediated interactions. The successful candidate will develop generalizable AI models that infer group‑level behavior from noisy, sparse, and imperfect real‑world multimodal data, including vision, audio, language, interaction traces, and contextual signals. This role is intended for an AI researcher who can make both fundamental and functional impact. The candidate should be able to advance core AI methods for multimodal learning, relational and temporal modeling, uncertainty‑aware inference, and human behavior modeling, while also translating research results into working prototypes and demonstrations. The work will support future systems that can understand group dynamics, social norms, roles, coordination, trust, engagement, alignment, conflict, and emergent collective behavior in complex human‑AI environments. The scientist will collaborate with interdisciplinary teams across AI, cognitive science, social behavior modeling, robotics, and human‑AI interaction to build AI systems that remain relevant as interaction patterns, sensing technologies, and AI‑enabled environments evolve over the next decade.

San Jose, CA

Key Responsibilities

• Develop AI models for collective state sensing and group‑level behavior modeling using real‑world multimodal data
• Create robust methods for learning from noisy, sparse, incomplete, and temporally dynamic human behavioral signals
• Model social roles, norms, relationships, influence, coordination, attention, trust, affect, and emergent group patterns in human‑human and human‑AI systems
• Apply and advance methods in multimodal representation learning, graph and relational learning, temporal modeling, foundation models, probabilistic reasoning, causal inference, and interpretable AI
• Design evaluation methods, benchmarks, annotations, and experimental protocols for validating collective behavior models with empirical data
• Develop uncertainty‑aware and interpretable models that support both automated sensing and human understanding of group dynamics
• Translate research results into proof‑of‑concept prototypes, internal demonstrations, and applied research directions
• Collaborate with researchers across AI, human behavior, cognitive modeling, social science, and human‑AI interaction
• Publish in leading research venues, contribute to intellectual property, and support HRI's long‑term research strategy

Minimum Qualifications

• Ph.D. in computer science, AI or a closely related field.
• Strong research record in highly reputable conferences and journals in AI/ML, multimodal learning, social signal processing, and human‑AI interaction.
• Strong research experience in multimodal signal processing with human data, including vision, audio, and behavioral signals under noisy or imperfect data conditions.
• Demonstrated ability to formulate open‑ended research problems, design rigorous experiments, and evaluate AI models on real‑world data.
• Experience developing and applying deep learning models, including both LLM‑based and non‑LLM architectures.
• Strong technical background in at least three of the following areas: multimodal learning, computer vision, speech/audio processing, natural language processing, graph or relational learning, temporal modeling, probabilistic modeling, causal inference, foundation models, or multi‑agent modeling.
• Strong programming skills in Python and experience with modern machine learning frameworks such as PyTorch, TensorFlow, or similar.
• Strong written and verbal communication skills, with the ability to present technical results to both expert and cross‑functional audiences.

Bonus Qualifications

• Track record of publications in relevant venues such as NeurIPS, ICML, ICLR, CVPR, ACL, EMNLP, CSCW, AAAI, IJCAI, AAMAS, and ICASSP.
• Experience conducting or analyzing human‑AI interaction studies.
• Familiarity with computational psychology, cognitive modeling, social behavior modeling, or related research areas.
• Experience with foundation models, vision‑language models, LLM‑based reasoning, graph neural networks, or multimodal representation learning.
• Experience developing experiments, running data collections, and conducting annotations in naturalistic or real‑world environments.
• Strong hands‑on experience translating research results into prototypes, demonstrations, software tools, or higher‑readiness applied research systems.
• Ability to work in interdisciplinary research settings and translate scientific ideas into prototype systems or applied research directions.

Desired Start Date

8/31/2026

Position Keywords

Multimodal AI, Social Signal Processing, Computer Vision, Human‑AI Group Interaction, Affective Computing, Foundation Models