About
About Me
I am an M.Eng. Robotics candidate at the University of Maryland with 2+ years of industry experience building autonomous systems, from ship hull inspection crawlers to legged robots. At GAMMA Lab, I research vision-language-action models that have achieved ~46% lower collision rates on real quadruped hardware.
My work sits at the intersection of perception, motion planning, and embedded systems. I have deployed software on Ghost Robotics Vision60, Unitree GO-2, TurtleBot4, and custom inspection platforms, integrating SLAM, Nav2, sensor fusion, and low-level C++ communication stacks (RS485, CAN, I2C) end-to-end.
I am an active ROS 2 open-source contributor (top 20, Kilted Kaiju release) and bring a hardware-first mindset to learning-based robotics. Currently seeking a Summer / Fall 2026 internship in autonomous systems, robot perception, or embedded robotics.
Skills & Expertise
Work Experience
Graduate Research Assistant
GAMMA Lab, University of Maryland
Built a human-in-the-loop annotation pipeline using OpenCV to generate 10,000+ counterfactual preference labels on the SCAND dataset for fine-tuning visuomotor navigation policies. Validated on Ghost Robotics Vision60 quadruped; policy achieved ~46% lower collision rate and ~24% higher goal-reach success in unstructured indoor environments. Contributing to 2 papers submitted to CoRL 2026: CHOP and VEGA.
Robotics Software Engineer Intern, SLAM & Quadruped Navigation
9thPixel Geosoft, Hyderabad, India
Designed wireless ROS 2 telemetry pipeline that cut operator command latency by ~60%, enabling stable remote operation at 100 m over UDP/Wi-Fi. Defined system architecture for a solar panel inspection robot (sensor placement, ROS 2 node topology, and hardware safety interlocks), reducing estimated hardware integration time by 30%.
Robotics Software Engineer, Embedded Systems & Autonomous Navigation
Octobotics Tech, Noida, India
Engineered autonomous magnetic-wheel crawler for ship hull inspection, tested at 2 international port sites; completed cycles 2× faster than manual procedures. Architected C++ RS485 libraries for STM32 actuators sustaining 10 Mbps throughput and 99.8% delivery reliability. Hardened communication stack with watchdog timers, cutting field downtime by ~70%. Built Qt5 teleoperation HMI that reduced operator onboarding from 2 weeks to 3 days.
Robotics Software Intern, Perception & UAV Systems
Pashaan Technologies, Nagpur, India
Developed OpenCV ArUco marker detection pipeline for autonomous VTOL precision landing, achieving <10 cm accuracy at 20 m altitude. Implemented SRT subtitle parsing pipeline cutting post-processing time by 65% across 200+ hours of footage. Engineered configurable Li-ion BMS supporting 3S–6S configurations with real-time balancing and safety cutoffs, deployed on 4 UAV platforms.
Education
Master of Engineering in Robotics Engineering
University of Maryland, College Park
GPA: 3.70/4.00 | Coursework: Motion Planning, Robot Perception, Autonomous Vehicles, Robot Learning, Swarm Robotics
Bachelor of Technology in Electronics Engineering
Government College of Engineering, Amravati
Minor in Telecommunication | CGPA: 7.65/10.00