GOTI
I'm a Robotics Engineer working on autonomous systems. How robots perceive their environment, estimate where they are, plan where to go, and execute reliably when things don't go as planned.
Currently a Robotics and AI Engineer Co-op at Piaggio Fast Forward in Boston, where I build simulation environments in NVIDIA Isaac Sim, develop perception and autonomy software in ROS2 and C++ to make mobile robots that move safely and intelligently in the environment. At Texas A&M, my work spans SLAM pipelines, RL based planners for drones and robotic arms, LiDAR camera fusion, and safe autonomous navigation. The thread connecting all of it is the same question: how do you build a system that holds up outside the lab?
Before grad school, I spent four years at NIT Karnataka building robots for ABU Robocon. I started as a programmer on the electronics team and finished as Vice Captain, leading a 40 person team through a full competition cycle. That experience taught me that the gap between a robot that works and one that doesn't is almost never the algorithm.
When I'm away from the lab, I'm usually on a trail. I trek, I photograph, and I document places. The same instinct to observe carefully and bring something back, just applied to mountains instead of sensor data.
- Building simulation environments in NVIDIA Isaac Sim that replicate real-world conditions, enabling the development and validation of autonomous navigation systems before physical deployment. Writing production-level robotics software in C++ and Python within the ROS2 Jazzy framework, contributing to perception and navigation pipelines that power mobile robots operating in dynamic, human-populated environments.
- Received Indian Academy of Science Summer Research Fellowship.
- Designed a PPO-based deep reinforcement learning trajectory planner for a 7-DOF robotic arm, enabling collision-aware and smooth goal-directed motion in continuous state-action spaces without relying on explicit motion planning models. Evaluated the learned policies in simulation across varied goal configurations, assessing performance through reward convergence curves, end-effector target accuracy, and trajectory smoothness metrics, demonstrating reliable and generalizable control behavior for high-DOF robotic manipulation.
Vice Captain — CSD Robocon NITK
