Research Scientist Intern, Dexterous Manipulation and Contact-Rich Control (PhD)

Agios is a full stack research organization within Reality Labs Research, spanning materials and manufacturing, sensors and actuators, simulation, and learned control. Our mission is to build systems that enhance human capabilities through intelligent machine collaboration. Today that work centers on what we call the last bottleneck: AI has compressed the digital side of research and development, while the physical side still moves at human speed. Dexterous Manipulation is the pillar attacking that problem directly, pairing a custom dexterous robotic hand with in-house wearable sensing, high accuracy simulation, and glove-driven teleoperation.

As a Research Scientist Intern, you will work on fast, stable, learned feedback control for multi-fingered robotic hands. Your focus will be reinforcement learning in high-fidelity simulation, tactile and proprioceptive feedback, and transfer to real robot hands, with mentors who span hardware, simulation, controls, and policy learning.

Our internships are sixteen (12) to twenty-four (24) weeks long and we have various start dates throughout the year.

Responsibilities

Drive an independent research thread on learned reactive control for dexterous hands, from problem formulation through simulation studies and real-robot evaluation.
Design controlled experiments and benchmarks that isolate issues with manipulation dexterity.
Integrate proprioceptive signals into control policies, and close the sim-to-real gap for contact through system identification, domain randomization, and realistic models of upstream policy error.
Collaborate across an interdisciplinary team of hardware, simulation, controls, and policy researchers, and communicate research plans, progress, and results.

Qualifications

Currently has or is in the process of obtaining a Ph.D. degree in Computer Science, Robotics, Mechanical or Electrical Engineering, or relevant technical field
Research experience in reinforcement learning or learning-based control for robotic manipulation or locomotion, demonstrated through publications, open-source code, or substantial projects
Experience with Python, and with a physics simulator such as MuJoCo, Isaac, or Drake
Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment Proven track record of achieving significant results as demonstrated by grants, fellowships, patents, as well as first-authored publications at leading workshops or conferences such as RSS, CoRL, ICRA, IROS, T-RO, IJRR, NeurIPS, ICML, or ICLR
Experience with sim-to-real transfer of contact-rich, locomotion, or dexterous manipulation policies, including domain randomization, system identification, and teacher–student distillation
Experience using tactile sensing, force/torque sensing, or contact estimation in policy learning or control
Experience in retargeting for locomotion or manipulation
Grounding in manipulation mechanics and control, such as contact modeling, impedance and admittance control, and trajectory optimization
Hands-on experience with multi-fingered robotic hands or grippers on real hardware
Experience working and communicating cross functionally in a team environment
Availability for minimum 16 consecutive week internship