AYUSH KUMAR

About Me

I’m a Robotics Engineer specializing in SLAM, navigation, computer vision, and reinforcement learning. With over 3 years of experience, I’ve contributed to the development of diverse autonomous systems and robotic platforms.

I am currently pursuing a Master’s degree in Robotics at Purdue University, with a focus on multi-modal models, reinforcement learning, and classical control theory to drive more efficient and intelligent robotic operations.

SLAM Reinforcement Learning Navigation Computer Vision Machine Learning ROS 2

Get In Touch Download Resume

My Experience

Featured Projects

Autonomy Stack for Quadruped - Addverb

Developed the autonomy stack—including SLAM and navigation—for Addverb’s in-house quadruped, TRAKR, enabling efficient navigation in unknown environments with minimal human intervention.

C++ Graph Optimization Bundle Adjustment Loop Closure Sensor Fusion Hybrid A* Pure Pursuit Controller Design Pattern OpenCV

VR Haptic Simulator - Addverb

Built a VR haptic simulator in Unity to train professionals in adapting to force feedback for long-distance teleoperation over 5G.

Unity C++ C# Interprocess communication 5G Haptic Rendering Spline Interpolation

Navigation among movable objects - IDEAS Lab

An intelligent navigation approach for dense and cluttered indoor and outdoor environments, utilizing the Boston Dynamics SPOT robot to interact effectively with its surroundings.

Isaac Lab VLM UNet Synthetic Data Generation Reinforcement Learning ROS 2 Python

AMR Software Stack - Orangewood Labs

Developed an extensive software stack—including navigation, mapping, localization, communication, data filtering, and kinematics—for an AMR equipped with a robotic arm for wall painting, using ROS 2 as the core framework.

Python C++ Nav2 ROS2 TCP/IP GMapping Firmware

6D DOF Pose tracker for Crane hooks

Developed a plug-and-play SMART BOX to track the 6-DoF pose of a crane’s hook, enhancing safety by enabling precise position and orientation tracking. Built as part of a robotics hackathon organised by L&T.

C++ Python Yolov4 RaspberryPi OpenCV Lidar ROS Kalman Filtering WebSocket

Sewer Monitoring Robot

Built a smart sewer monitoring robot that maps underground pipelines and performs structural analysis, enabling efficient inspection, early fault detection, and data-driven maintenance planning. Built as a part of hackathon organised by La Trobe University Australia.

C++ Python SPI/I2C Serial RaspberryPi Yolov4 OpenCV SLAM A* Sensor Fusion

UV disinfectant robot

Designed and implemented an autonomous UV-disinfection robot for hospitals, featuring a custom algorithm that intelligently optimizes zap patterns to reduce power usage and maximize area coverage. Completed as part of an undergrad's capstone project.

C++ Python SPI/I2C Serial Data Structures RaspberryPi Yolov4 OpenCV SLAM A* Sensor Fusion

Efficient Ransac based ground plane segmentation algorithm using IMU

Designed and implemented a RANSAC-based ground plane segmentation algorithm, using IMU readings as an initial estimate to reduce computation time and improve real-time performance.

C++ Realsense OpenCV RANSAC IMU SVD

Mechatronic Rehabilition Glove

Developed a mechatronic rehabilitation glove aimed at supporting motor recovery in children from refugee camps in India. Integrated with a game to reduce the physical and emotional strain of therapy by making the process playful and engaging. Built as a part of government project.

Python C++ Polynomial Regression Arduino Firmware EMG TCP/IP Interprocess Communication Lua CoppeliaSim

6 DOF robotic ARM

Engineered a fully functional 6-DoF robotic arm from scratch, writing low-level firmware for actuators and implementing ROS-based control architectures. The project included testing various motion planning strategies to enable precise and adaptive manipulation.

Python C++ MoveIt ROS TCP/IP CAN Firmware Velocity Controller

Custom Path Planning Algorithm

Developed a lightweight, ROS-independent A* path planning algorithm with path smoothing, fully implemented in C++ for deployment on resource-constrained boards.

C++ Design Pattern Data Interpolation Priority Queue Custom Map Datatype

Motion Capture Robot

Designed a humanoid robot controlled via a sensor-equipped wearable suit, enabling real-time imitation of human movements. Developed as part of a national-level hackathon project during high school.

Arduino C++ Python C# ROS Analog Signal Analysis Servo Control

Publications

• A Sensor-based system for real-time three-dimensional positioning of crane hook and method thereof,

  Ayush Kumar, Co-author • published by Intellectual Property India, 2023

  Application No. 202321075984

• Haptic device for six degrees of freedom parallel manipulation

  Ayush Kumar, Co-author • published by Intellectual Property India, 2023

  Application No. 202311074914

• Design of mechatronic rehabilitation glove

  Ayush Kumar, Co-author • published by Intellectual Property India, 2022

  Application No. 361771-001

Achievements

Gold Medal BTech - 2024

Awarded the D&M Proficiency Gold Medal for the best project across the entire B.Tech batch of 2023.

Rookie Of The Year - 2024

Nominated as Best Performer among all freshers in the Advanced Robotics division at Addverb.

Top 10 in TIGC - 2024

Ranked in the top 10 at TIGC, an international robotics hackathon organized by La Trobe University, Australia, and received AUD 1,000 (~USD 670) in funding to develop a prototype of the proposed solution.

AIR-1 Techgium - 2022

Secured AIR-1 among 800+ participants in a hackathon organized by Larsen & Toubro (L&T), winning a cash prize of INR 1M (~11K USD ) for developing a solution to mitigate crane accidents at construction sites.

Winner Technovate - 2020

Awarded first place in Technovate, a national-level exhibition at IIIT Jabalpur during the annual techno-cultural and design festival.

Winner BullsEye - 2020

Secured 1st place in BullsEye, a national-level rocket design competition at IIIT Jabalpur’s annual techno-cultural and design festival, earning titles for both longest flight time and maximum range.

RunnerUp Robothon - 2020

Finished as the runner-up in Robothon, a national-level all-terrain semi autonomous vehicle design competition at IIIT Jabalpur during the annual techno-cultural and design festival.

Winner National-level CBSE Science Exhibition - 2017

Won the CBSE National Level Science Exhibition for developing a motion-synchronized humanoid robot torso that mimics the movements of a human operator.