Research Scientist Intern, Robot Foundation Models (PhD)

About the Role

Agios, a full-stack research organization within Reality Labs Research, is focused on enhancing human capabilities through intelligent machine collaboration. The Dexterous Manipulation pillar aims to accelerate the physical side of research and development by pairing a custom dexterous robotic hand with advanced sensing, simulation, and teleoperation technologies. As a Research Scientist Intern, you will work at the intersection of robot learning, dexterous control, and human-to-robot transfer, contributing to learned models that translate human demonstration data into executable robot behavior.

Responsibilities

• Drive an independent research thread on learned control and motion transfer for dexterous hands, from problem formulation to evaluation in real robots or high-fidelity simulations.
• Build training pipelines for modern policy classes, including flow matching, diffusion policies, imitation learning, reinforcement learning, large behavior models, and VLAs.
• Develop representations that fuse multimodal human and robot data (video, hand pose, tactile/contact force, IMU) into a shared action space.
• Construct data collection pipelines and benchmarks for dexterous manipulation, including protocols for generalization across objects, motions, and embodiments.
• Collaborate with an interdisciplinary team of hardware, simulation, controls, and policy researchers, and communicate research plans, progress, and results.

Requirements

• Understanding of the state of the art in learned dexterous manipulation and control, with judgment about which open problems are worth advancing.
• Experience with Python and PyTorch or JAX, plus familiarity with ROS or comparable robotics middleware.
• Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment.

Qualifications

• Proven track record of significant results, as demonstrated by grants, fellowships, patents, and first-authored publications at leading conferences (RSS, ICRA, IROS, CoRL, NeurIPS, ICML, etc.).
• Experience training modern control policies (flow matching, diffusion, imitation learning, reinforcement learning, large behavior models).
• Familiarity with robotics modeling and analysis (kinematics, dynamics, trajectory optimization, motion planning, perception, task planning, control theory).
• Hands-on experience with multi-fingered robotic hand manipulation and parallel-jaw grippers mounted on robot arms.
• Experience with motion retargeting, sim-to-real workflows, and real robot demonstration collection (teleoperation, kinesthetic guidance, wearable capture).
• Experience with multimodal representation learning, vision-based input recognition, and time sequence sensor data (tactile, IMU).
• Experience designing data collection protocols and building diverse datasets.
• Experience with prototype software and hardware systems (robot platforms, multi-camera rigs, wearable sensing).
• Availability for a minimum 16-week consecutive internship.

Pay

$7,650/month to $12,667/month + benefits. Individual compensation is determined by skills, qualifications, experience, and location. Compensation details listed reflect the base rate only and do not include bonuses, equity, or incentives.

Schedule

Internships are twelve (12) to twenty-four (24) weeks long, with various start dates throughout the year.