Research Assistant — Robotic Manipulation Simulation & Sim-to-Real Transfer

Stevens Institute of Technology•Hoboken, NJ

9d•Onsite

About The Position

The Advanced Robot Manipulators Lab at Stevens Institute of Technology, led by Prof. Long Wang, is seeking a PhD-level Part-Time Research Assistant to lead simulation-driven research on dexterous robotic manipulation and sim-to-real transfer. The Research Assistant will build high-fidelity manipulation simulations in NVIDIA Isaac Sim / Isaac Lab and transfer the resulting model-based and learning-based policies onto physical robotic hardware. Working alongside faculty, graduate students, and industry partners on Department of Defense–funded research, the Research Assistant will drive the full pipeline from simulation design through hardware validation, close the sim-to-real gap on real manipulators, and contribute to sponsor deliverables. This position offers a doctoral-level researcher deep, hands-on involvement in state-of-the-art robotics research and direct exposure to industry-academic collaboration.

Requirements

PhD in Robotics, Mechanical Engineering, Electrical Engineering, Computer Science, or a closely related field (earned, or with degree requirements substantially complete).
Demonstrated research experience in robotic manipulation, reinforcement/imitation learning, or model-based control.
Hands-on experience with physics-based robot simulation, especially NVIDIA Isaac Sim / Isaac Lab (MuJoCo, PyBullet, or comparable also considered).
Proficiency in Python, with familiarity in PyTorch or a comparable machine-learning framework.
Experience with sim-to-real transfer techniques such as domain randomization, system identification, or calibration.
Experience deploying and validating algorithms on physical robot hardware.
Strong record of research output (e.g., peer-reviewed publications) and clear written and verbal communication.

Nice To Haves

Experience with ROS / ROS 2 and real-time robot control.
Familiarity with dexterous robotic hands (e.g., LEAP hand variants) and tactile, force/torque, or vision-based sensing.
GPU-accelerated, large-scale parallel simulation and training (Isaac Lab / Orbit).
USD / OpenUSD asset authoring for simulation environments.
Experience mentoring students and leading collaborative research efforts.

Responsibilities

Design, build, and validate high-fidelity robotic manipulation simulations in NVIDIA Isaac Sim / Isaac Lab, including scene construction, contact and physics tuning, and sensor modeling.
Develop, train, and benchmark manipulation policies (model-based control and/or reinforcement and imitation learning) for dexterous hands and manipulators.
Design and execute sim-to-real transfer pipelines, including domain randomization, system identification, and calibration to reduce the sim-to-real gap.
Deploy simulated policies onto physical robot hardware, quantify sim-to-real performance gaps, and iterate on both simulation and hardware.
Integrate perception, tactile/force sensing, and control stacks so that they behave consistently across simulation and hardware.
Maintain reproducible codebases, data-acquisition workflows, and technical documentation.
Contribute to design reviews, technical reports, and progress updates for sponsor deliverables, and help mentor junior graduate and undergraduate students.
Collaborate with faculty, PhD students, and external industry and university partners on shared research goals.