Intern, Multimodal Embodied AI & 3D Scene Design

Mitsubishi Electric Research Laboratories, Inc.; Cambridge (MA); On-site; USD 42,000 - 60,000; Full time

Job summary

Mitsubishi Electric Research Labs, Inc. (MERL) invites a self-motivated intern to explore multimodal large language models and embodied AI in dynamic indoor environments.

The candidate should be a PhD student with a strong ML/CV background and top-tier publications, with experience in synthetic 3D scenes, simulators like Habitat/SoundSpaces, and ML/RL methods. Hands-on work includes collaborating with MERL CV researchers and preparing manuscripts for scientific venues.

Qualifications

• PhD student with top-tier publications in ML/CV.
• Experience designing synthetic 3D scenes using graphics software.
• Experience in vision-based embodied AI and simulators.
• Experience training large language models on multimodal data.
• Background in reinforcement learning.
• Strong ML fundamentals and solid programming skills.

Responsibilities

• Collaborate with MERL researchers to develop algorithms.
• Prepare manuscripts for scientific publications.
• Advance multimodal LLM and embodied AI in dynamic indoor settings.
• Design 3D interactive scenes and evaluate in simulators.

Skills

Machine learning

Computer vision

Reinforcement learning

Multimodal learning

Programming

Publications

Education

PhD student

Tools

Habitat

SoundSpaces

SMPL

Job description

Mitsubishi Electric Research Labs, Inc. (MERL) invites a self-motivated intern to explore multimodal large language models and embodied AI in dynamic indoor environments.

The candidate should be a PhD student with a strong ML/CV background and top-tier publications, with experience in synthetic 3D scenes, simulators like Habitat/SoundSpaces, and ML/RL methods. Hands-on work includes collaborating with MERL CV researchers and preparing manuscripts for scientific venues.