Experience

Where I’ve worked, what I worked on, and what I know how to use.

Professional

  • Sept 2025 – Present

    Robotics 3D Specialist

    Arrive AI Inc. · Fishers, Indiana

    • Building a digital twin simulation environment at scale in NVIDIA Isaac Sim for scenario testing before real-world deployment.
    • Developing the autonomy architecture for a mobile robot platform, integrating reinforcement learning policies with vision-language (VLM) and vision-language-action (VLA) models — including deploying open VLA models such as OpenVLA and OpenPI — for instruction-driven tasks.
    • Evaluating and optimizing vision transformers for real-time perception, accelerating inference with NVIDIA Triton, the DeepStream SDK and TensorRT.
    • Profiling edge devices for inference latency and throughput trade-offs — Jetson Orin Nano, Orin AGX and Thor (Blackwell), plus Qualcomm IQ9075 and Infineon Technologies hardware.
    • Architecting a containerized AI inference codebase with Docker for repeatable deployment, working with the platform team on k3s fleet rollout, and building the web frontend for low-latency WebRTC video streaming.

  • Feb – July 2021

    Controls Engineer

    Caterpillar Inc. · Peoria, Illinois

    • Developed sub-system level control for large power engine systems with the engine performance simulation team.
    • Tested engine component models for thermal anomalies and tuned gain scheduling using GT-Power, MATLAB and Simulink.
    • Restructured 10+ engine sub-system modules in Simulink, improving model readability and fault diagnosis.

Research

  • Aug 2021 – Sept 2025

    Graduate Research Assistant

    ARM Lab, Clemson University (CU-ICAR) · Greenville, South Carolina

    • Dissertation: Hybrid Learning for Rough Terrain Navigation of Actively Articulated Wheeled Vehicles.
    • Coordinated payload transport with biped-wheeled robots — a single DRL agent controlling two robots, transferred zero-shot to hardware (project, ICRA 2025).
    • Hybrid deep reinforcement learning for rough-terrain path tracking, validated to 98.5% precision on unseen real tracks (project, AIM 2024).
    • Agile off-road terrain traversal with learned goal-to-goal navigation, 86% goal success (project, IFAC MECC).
    • Real-time 3D scene reconstruction with NVIDIA Isaac ROS Nvblox, in collaboration with the NVIDIA Robotics team (project).
    • Operationalized Isaac Bi3D freespace segmentation on the Palmetto HPC cluster and on a Clearpath Husky (project).
    • LiDAR-inertial odometry and mapping on a Unitree Go1 quadruped using LIO-SAM (project).

Education

  • Graduated Sept 2025

    PhD, Automotive Engineering

    Clemson University · Greenville, South Carolina

    • Dissertation: Hybrid Learning for Rough Terrain Navigation of Actively Articulated Wheeled Vehicles.

  • Graduated May 2020

    MS, Mechanical Engineering

    Michigan Technological University · Houghton, Michigan

  • Graduated May 2017

    BE, Mechanical Engineering

    University of Mumbai · Mumbai, India

Skills

Focus areas
Reinforcement learning, perception, autonomy and navigation, vision-language (VLM) and vision-language-action (VLA) models, robotics simulation platforms, digital twins, simulation-in-the-loop testing and evaluation, sim-to-real / zero-shot transfer.
Simulation and 3D
NVIDIA Isaac Sim, Isaac Lab, Omniverse, OpenUSD, PhysX, Blender, Gazebo, URDF.
Frameworks and tools
ROS, ROS 2, PyTorch, ONNX, TensorRT, Triton, DeepStream SDK, OpenVLA, OpenPI, WebRTC, Docker, Git, MATLAB, Simulink.
Hardware and platforms
Unitree Go1 quadruped, Clearpath Husky, TurtleBot3, Jetson Orin Nano / AGX / Thor (Blackwell), Qualcomm IQ9075, Infineon Technologies hardware, Velodyne VLP-16 LiDAR, Intel RealSense and stereo depth cameras, OptiTrack motion capture.
Languages
Python, C++.

Want this as a document? Ask me and I'll send the current version.