Meta
Research Scientist Intern, Multimodal Learning for Robotics (PhD)
Burlingame, CA · Internship · Posted Oct 21, 2026
- Experience
- No years specified.
- Education
- Currently has or is in the process of obtaining a Ph.D. degree in Computer Science, Artificial Intelligence, Robotics, or relevant technical field.
- Visa sponsorship
- Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment.
About the job
Develops computer-vision and tactile algorithms, data pipelines, and models for robotic dexterous manipulation, while collaborating on research and communicating results.
What you'll do
- Develop CV- and tactile-based algorithms for human and robotics data.
- Build robotic data pipelines, train models, and benchmark algorithms for dexterous manipulation.
- Collaborate with researchers and cross-functional partners; communicate research plans, progress, and results.
What they ask for
- Currently has or is pursuing a Ph.D. in Computer Science, Artificial Intelligence, Robotics, or a relevant technical field.
- Experience in computer vision, with active contributions in at least one of 3D reconstruction, hand pose estimation, or object pose estimation, and strong familiarity across all.
- Understanding of state-of-the-art robotic manipulation control policies and experience with physics-based simulators like Mujoco.
- Experience with Python, PyTorch or JAX, or related languages; must obtain and maintain work authorization in the country of employment.
Nice to have
- Experience with multimodal learning, including tactile data, and robot manipulation.
- Experience building pipelines with foundation CV models such as CLIP, DinoV2, or SAM2.
Pay
$7,650/month to $12,667/month (base salary)
Tools
Mujoco, Python, PyTorch, JAX, CLIP, DinoV2, SAM2
AI focus
Applies computer vision, tactile sensing, and multimodal learning to robotic control and dexterous manipulation.
Model-written summary · the original posting is authoritative