Sugheerth Sreedharan

I am a PhD student in Computer Science at the University at Buffalo, advised by Dr. Karthik Dantu in DRONES Lab. My research focuses on safe motion planning, diffusion-based planning, trajectory generation, and sim-to-real transfer for highly articulated robotic manipulators and autonomous systems.

I work on autonomy stacks that connect planning, control, learned dynamics, simulation, and field deployment. My ongoing research includes trajectory-level constraint-aware safe sampling for diffusion models, diffusion-based trajectory planning for excavators with learned forward and inverse dynamics models, and dynamics-guided sampling for physically coherent observation-control trajectories.

In the EARTH project, I contribute to safe planning and control, ROS2 lifecycle systems, Nav2-based navigation, MoveIt2 arm planning, CAN bus integration, and deployment on hydraulic excavators. I am also developing methods that use learned hydraulic dynamics to generate demonstrations in simulation and guide diffusion planners toward trajectories that are feasible for real excavator hardware.

Before graduate research, I worked as a software development engineer at Zoho Corporation, building cloud repository management systems, search infrastructure, gRPC services, and database-backed backend platforms.