Shubham Shah

About

My name is Shubham Shah, and I am currently pursuing an MS in Robotics, with a focus on localization and motion planning of robots. Prior to this, I worked as a robotics engineer at NewSpace Research and Technologies, where I designed and implemented motion planning and Localization systems for swarms of UAVs in complex GPS-denied environments, specifically for emergency response and disaster management. I completed my B-Tech in Mechatronics in 2020, and since then, I have gained extensive experience in computer vision, robotics, and motion planning. As a researcher, I am passionate about exploring both learning and non-learning-based methods in motion planning, robotics, and computer vision, striving to develop innovative solutions that drive real-world impact.

Resume

Proficient in Computer Vision, Robotics, and Motion Planning techniques with close to 4 years of industry experience. for more details, please download my detailed resume here.

Education

Master of Science in Robotics

2024 - 2026

University of Delaware, Delaware USA

exploring learning and non learning methods for decentralised localization coupled with motion planning

Education

Bachelor of Technology in Mechatronics

2016 - 2020

Mukesh patel school of technology management and engineering (NMIMS), India

Accquired Strong foundational knowledge in programming, DSA, control systems, robotics fundamentals, AI, and numerical optimization techniques

Open Source Projects

Contributions

I was able to contribute to the open-source community by actively participating in various projects. my contributions include, features addition, optimizing performance, code review, fixing and reporting bugs, and enhancing documentation to make it more accessible and comprehensive. Below I have mentioned a few projects I contributed to

Professional Experience

Robotics Engineer

May 2021 - Present

NewSpace Research and Technologies, Bengaluru, India

  • Check out our latest work on swarm navigation in highly dense forests here
  • Responsible for single and multi-agent and path planning and localization stack for UAVs in GPS‑Denied, dynamic, and obstacle-rich environments
  • Contributed to increasing the speed of UAVs, enhancing memory and CPU performance, reducing tracking error, and bolstering overall stability
  • Designed and incorporated functionalities including recovery behaviors, coordinated flight support, terrain navigation, and fallback mechanism
  • Built the complete pipeline for GPS-denied UAV-swarm navigation in simulation
  • Developed solutions for glass detection, dust filtering, collabrative mapping, recovery, sensor fusion, and localization drift compensation

Product Developer

July 2020 - May 2021

Om Siddh Vinayak Impex Pvt. Ltd., Gandhidham, India

  • Built a Face Recognition Pipeline for attendance system with modules for detection, tracking, head pose estimation from RGB image, and recognition, achieved an overall accuracy of 91%.
  • Implemented a Vision-based License plate recognition stack with modules for detecting, tracking, and OCR, developed a novel approach for OCR which reduced the recognition time by 40-45%. achieved an overall accuracy of 82%
  • Worked on model quantization, Pruning, and other optimization techniques to deploy the models on edge devices

Portfolio

Throughout my academic and professional experience, I have had the opportunity to work on numerous exciting and innovative projects. Here's a brief overview of some of them.

Swarm Forest Navigation

Developed path and motion planning stack for swarm UAV navigation in dense, cluttered enviroment like a forest

Forest Navigation

Developed path and motion planning stack for single UAV navigation in dense, cluttered enviroment like a forest

GPS-Denied Localization

Contributed to the development of UAV Localization techniques using lidar and inertial sensors in densely forested areas

Dust Filtering

Worked on dust filtering solution for generating clear 3D maps and aiding navigation

Facial recognition attendance system

Developed a facial recognition system to reduce attendance time in large scale system

3D Terrain Estimation and Navigation

Developed a soultion to Estimate the terrain gradient in real time and aid navigation

Smart Collection Box

Worked on automating and improving existing collection box system

Number Plate Recognition System

Developed a numberplate recognition system for vehicle attendance at large scale

Blogs

Authored over 40 articles covering algorithms, data structures, and key programming concepts.

Versatile Robotic Arm

Built a AI and IoT enabled Versatile Robotic arm to solve common warehousing issues, more details can be found in the report attached to the project

Contact

Let's connect! Feel free to drop me a mail

Location:

Delaware, USA