Internship, Multimodal Embodied AI
Mitsubishi Electric Research Laboratories
About the Role
MERL is seeking a self-motivated intern to address problems at the intersection of multimodal large language models and embodied AI within dynamic indoor environments.
Key Responsibilities
- Collaborate with researchers in computer vision at MERL.
- Develop algorithms for multimodal embodied AI.
- Prepare manuscripts for scientific publications.
Requirements & Tech Stack
- Prior experience in designing synthetic scenes (e.g., 3D games) using popular graphics software.
- Experience with embodied AI.
- Experience with large language models.
- Experience with reinforcement learning.
- Experience using simulators such as Habitat/SoundSpaces.
- Hands-on experience using animated 3D human shape models (e.g., SMPL and variants) is desired.
Qualifications
- PhD student.
- Strong background in machine learning.
- Strong background in computer vision.
- Demonstrated by top-tier publications.
- Experience in designing 3D interactive scenes.
- Experience with vision-based embodied AI using simulators.
- Experience training large language models on multimodal data.
- Experience with training reinforcement learning algorithms.
- Strong foundations in machine learning.
- Strong foundations in programming.
- Strong track record of publications in top-tier computer vision and machine learning venues (e.g., CVPR, NeurIPS).