Beth Israel Deaconess Medical Center (BIDMC)
Machine Learning Intern
Boston, MA
Sept 2025 – Present
- ·Build a GPU-enabled surgical robotics simulation workflow for the da Vinci Research Kit (dVRK) using ROS and Gazebo/AMBF on BU's Shared Computing Cluster (SCC) by containerizing dependencies with Singularity/Apptainer and creating reliable scripts to launch and view simulations remotely
- ·Prototype incision-closure tasks in NVIDIA Isaac Sim/ORBIT-Surgical by setting up a reproducible Omniverse/Isaac environment (fixed versions and dependencies) and organizing the required simulation assets for surgical manipulation
- ·Improve ML autonomy training reliability by diagnosing a preprocessing mismatch that reduced performance from ~90% to ~30%, curating/labeling demonstration data (e.g. ~80 JIGSAWS trajectories), and collaborating with Johns Hopkins to acquire new dVRK kinematic demos to standardize multimodal inputs (vision + kinematics)
PythonROSGazebo/AMBFNVIDIA Isaac SimSingularity/ApptainerdVRK