Mitsubishi-posted 14 days ago
$6,000 - $8,000/Yr
Intern
Onsite • Cambridge, MA
5,001-10,000 employees

MERL is seeking a highly motivated PhD student to conduct research on bimanual visuotactile manipulation for industrial applications, such as assembly, disassembly, and tool-enabled operations. The focus of this internship is on developing closed-loop manipulation skills for contact-rich tasks using visuotactile sensing and multi-modal learning from large foundation models. In particular, we are interested in exploring how VLA/RT-style vision-language-action models or foundation manipulation models can be adapted or fine-tuned for real-world bimanual settings. A successful internship will lead to a submission to a top-tier robotics venue (e.g., RSS, ICRA, CoRL) in collaboration with MERL researchers. The internship duration is approximately 3 months, beginning in Summer or Fall 2026, and will be fully onsite at MERL.

  • Senior PhD student in robotics, mechanical/electrical engineering, computer science, or related fields.
  • Machine learning for robotic manipulation (e.g., RL, imitation learning, representation learning, VLA/foundation model fine-tuning), with an emphasis on real-world systems.
  • Experience with visuotactile sensors (e.g., GelSight, GelSlim, DIGIT).
  • Experience working with physical robotic hardware.
  • Prior publication in top robotics venues (RSS, CoRL, ICRA).
  • Strong programming experience in Python and/or C++, especially with ML or control frameworks (e.g., PyTorch, MuJoCo).
  • Experience with foundation models for robotics, such as Vision-Language-Action models (e.g., RT-2, RT-X, OpenVLA), Multi-modal diffusion or transformer policies, Fine-tuning pipelines (e.g., LoRA, RLF).
  • Experience building high-quality robotics codebases (OOP, CI, unittests).
© 2024 Teal Labs, Inc
Privacy PolicyTerms of Service