Embodied AI & Humanoid Robotics Research Internship

merl; Cambridge (MA); On-site; USD 67,000 - 89,000; Part time

Benefits offered by this job

Relocation stipend

Travel covered to and from MERL

Charlie Card for local commuting

Health insurance after 90 days

Immigration support

Weekly research talks

Job summary

Join our cutting-edge research team at MERL to help advance Embodied AI and Humanoid Robotics. As a research intern, you will develop AI technologies enabling humanoid robots to understand, reason, and interact with the physical world through complex manipulation and tool-use tasks, with a goal of publishing at premier AI and robotics conferences.

Projects include embodied AI for manipulation, VLA models, long-horizon planning with WAM, and learning from demonstrations, teleoperation, and

Qualifications

• Strong research experience in robotics, embodied AI, ML, CV or related fields.
• Experience with deep learning frameworks such as PyTorch or JAX, and strong Python programming skills.
• Familiarity with Vision-Language-Action models, Foundation Models or multimodal AI, RL or imitation learning, robot manipulation, motion planning or control, and agentic AI for robotics.

Responsibilities

• Embodied AI for dexterous manipulation, assembly, and tool use.
• Vision-Language-Action models and Foundation Models for robotic control.
• World-Action Models for long-horizon planning and decision making.
• Learning from human demonstrations, teleoperation, and autonomous data collection.
• Sim-to-real transfer, reinforcement learning, and real-world robot deployment.

Skills

Robotics research

PyTorch

Python

Tools

ROS/ROS 2

Isaac Sim

MuJoCo

Genesis

Jetson GPUs

Job description

Join our cutting-edge research team at MERL to help advance Embodied AI and Humanoid Robotics. As a research intern, you will develop AI technologies enabling humanoid robots to understand, reason, and interact with the physical world through complex manipulation and tool-use tasks, with a goal of publishing at premier AI and robotics conferences.

Projects include embodied AI for manipulation, VLA models, long-horizon planning with WAM, and learning from demonstrations, teleoperation, and