Robotics Developer I at Delhivery Limited (2025-10 – Present)
- Building the autonomy stack for a tricycle-drive pallet truck AMR — covering the navigation layer (Nav2/behavior trees) and the hardware abstraction layer (ros2_control).
- Developed a custom Nav2 navigator (ROS 2 Humble) for route-server-based, node-to-node navigation: implemented a custom behavior tree that sequences route-server waypoint requests, planner/controller execution, and goal-reached checks, with a dedicated obstacle-detection condition node that halts the robot in place on obstacle intrusion instead of triggering the default recovery subtree.
- Implemented the Regulated Pure Pursuit (RPP) local controller for route-server-based path tracking.
- Built the ros2_control hardware interface as a custom resource manager plugin, integrating a UDP-based API for production and a serial API for prototype servo control.
- Built a custom tricycle ros2_controller plugin to calculate odometry, with an implementation of inverse and forward kinematics.
- Built a DAQ ros2_controller plugin that aggregates raw sensor data and republishes it in a structured form consumed by the higher-level state machine / orchestrator module — enabling end-to-end data transmission between the low-level controller and upper-level systems, including upper-to-low-level HMI state data flow, giving structured visibility into system state for diagnostics.
- Built a complete simulation environment (URDF, Gazebo, ros2_control, Cartographer SLAM, full Nav2 stack) to validate the above end-to-end.
Consultant Engineer at Flipkart Internet Private Limited (2024-12 – 2025-10)
- Designed the complete hardware architecture for retrofitting a manually operated pallet truck into an autonomous AMR — defined the STM32-based low-level control layer and interfaced it with a Curtis 1232E motor controller for traction, replacing manual throttle/steering input with electronically commanded control.
- Delivered end-to-end development of a free-space-based navigation simulation stack for the AMR: authored URDF kinematic/collision models in SolidWorks, built a Gazebo simulation environment with realistic physics and collision response, integrated Cartographer SLAM with 3D LiDAR for mapping and localization, configured the Nav2 stack (costmaps, planner/controller plugins) for free-space path planning, and implemented ros2_controllers for closed-loop simulation control.
- Gained hands-on exposure to a magnetic-tape-guided AGV's control architecture — motor controllers, safety LiDARs, RFID-based station/waypoint detection, and magnetic guidance sensors — understanding how a rule-based line-following AGV differs architecturally from a free-space SLAM/Nav2-based AMR.
Robotics Engineer Intern at Entudio Pvt Ltd (2023-01 – 2023-06)
- Conducted hands-on multirotor workshops, guided engineering students on final-year projects, and operated 3D printers to design and fabricate custom parts.
