Kartik Virmani
Robotics Engineer & Research Assistant
MS in Mechanical Engineering & Applied Mechanics at the University of Pennsylvania, specializing in Mechatronics, Robotics, and Autonomous Systems.
MS in Mechanical Engineering & Applied Mechanics at the University of Pennsylvania, specializing in Mechatronics, Robotics, and Autonomous Systems.
I'm a passionate Robotics Engineer currently pursuing my Master's at the University of Pennsylvania. With a strong foundation in mechanical engineering and hands-on autonomous systems work, I build full-stack robotics solutions—from embedded control and estimation to motion planning and perception.
My research focuses on mobile humanoid robots, autonomous racing, and advanced control systems. I work daily with ROS 2 (C++/Python), Gazebo/RViz, and modern planning & learning methods.
University of Pennsylvania • May 2026
Concentration: Mechatronics & Robotics
VIT — Vellore Institute of Technology • May 2022
GPA: 9.73/10.00
Full-stack ROS 2 autonomy integrating localization, navigation, and motion planning (PRM/A*, Pure Pursuit, MPC) with RealSense-based mapping.
3D planning (A*/OMPL), VIO/SLAM-based localization, and PID/MPC control in a physics-accurate simulator with hardware-in-the-loop tests.
GroundingDINO + SAM + Kalman smoothing for ORB-SLAM3; reduced ATE and boosted map density significantly on dynamic scenes.
SVGD-guided MPPI on F1/10 at 50 Hz; reduced max track error and cross-track error with stable high-frequency control.
Random-shooting MPC and PPO (Walker/Humanoid) with stable convergence and strong returns on DeepMind Control tasks.
Fused IMU, GPS, LiDAR, and encoders via UKF; demonstrated reduced localization drift over kilometer-scale routes.
PyTorch NeRF with positional encoding and hierarchical sampling; evaluated with PSNR/SSIM on novel-view synthesis.
CarSim + MATLAB + CUDA to hybridize NN dynamics with MPC; cut planning convergence time substantially.
Zero-Momentum Control + MPC in PyDrake/MuJoCo; improved planning robustness across disturbance tests.
ESP32 modular platform with ToF, H-Bridge, encoder PID loops, and Wi-Fi multi-MCU comms for indoor autonomy.
Transient RANS in STAR-CCM+ to optimize undertray & diffuser aerodynamics across pitch angles.
SolidWorks + ANSYS structural validation under ≤700 N steering loads with GD&T-driven tolerances.
Thermo-mechanical FEA for heat/wear modeling; geometry + materials tuned for fatigue resistance.