top of page
My Portfolio
Welcome to my portfolio. Here you’ll find a selection of my work. Explore my projects to learn more about what I do.
ECE 484 Final Project
Developed the software for an autonomous F1-Tenth vehicle, focusing on mapping and navigation. We created a system that could detect road signs with high accuracy, ensuring the vehicle followed road rules
Developed the software for autonomous mapping of an F1-Tenth vehicle using ROS2, Jetson, and over 900 lines of Python code for real-time navigation and mapping.
Implemented YOLOv5 and Vicon for 97% accurate road sign detection and mapping, ensuring reliable adherence to traffic signs.
Fused map data with A search algorithms* for real-time shortest path navigation, achieving 85% task completion through accurate route optimization.
Utilized Monte Carlo Localization to verify vehicle position, enhancing the accuracy and performance of autonomous driving.
Engineered real-time lane detection using computer vision (OpenCV) techniques, adjusting for variable lighting conditions, and incorporating a PID controller to refine steering angles and enhance path accuracy.
Developed the software for autonomous mapping of an F1-Tenth vehicle using ROS2, Jetson, and over 900 lines of Python code for real-time navigation and mapping.
Implemented YOLOv5 and Vicon for 97% accurate road sign detection and mapping, ensuring reliable adherence to traffic signs.
Fused map data with A search algorithms* for real-time shortest path navigation, achieving 85% task completion through accurate route optimization.
Utilized Monte Carlo Localization to verify vehicle position, enhancing the accuracy and performance of autonomous driving.
Engineered real-time lane detection using computer vision (OpenCV) techniques, adjusting for variable lighting conditions, and incorporating a PID controller to refine steering angles and enhance path accuracy.
Robosub Robotics Submarine
As the Captain of the Autonomous Robosub team at iRobotics, I led the revitalization of a 15-person robotics submarine team competing in the RobotNation Robosub competition—an autonomous underwater robotics challenge involving complex underwater tasks.
Key accomplishments include:
• Modular Design Implementation: Introduced a modular design approach that reduced the submarine's weight by 47%, improved heat dissipation, and made it highly serviceable/debuggable
• Debugging old 2 year old PCB from scratch, identified 4 redundant parts and replaced them to make an efficient and modular PCB design
•Facilitated a remote thruster control pipeline from Nvidia Jetson using ROS, Cube controller and VNC viewer
Key accomplishments include:
• Modular Design Implementation: Introduced a modular design approach that reduced the submarine's weight by 47%, improved heat dissipation, and made it highly serviceable/debuggable
• Debugging old 2 year old PCB from scratch, identified 4 redundant parts and replaced them to make an efficient and modular PCB design
•Facilitated a remote thruster control pipeline from Nvidia Jetson using ROS, Cube controller and VNC viewer
Harvard Autonomous Pacbot Competition
Engineered an RTOS autonomous robot that has won and set the record at the Harvard Pacbot Autonomous Robotics Challenge, navigating a maze-like arena, collecting "points" while avoiding or attacking "ghost" robots
•Achieved historical victory by becoming the first team to completely clear the maze
•Facilitated the development of a custom 2-sided custom PCB to accommodate multiple connections in a small space
•Identified bottlenecks and established a streamlined and 35% more efficient calibration and communication pipeline
•Implemented innovative solutions across hardware and software, integrating a Raspberry Pi, Teensy, 9 sensors, and motors into a cohesive unit. Authored and debugged over 2,000 lines of C++ code, ensuring precise control and seamless interaction with the Pacbot environment.
•Developed and troubleshooted a gesture control system using ESPNOW, IMU Sensors and C++, creating an interactive robotics experience for over 25,000 visitors at the Engineering Open House, an engineering showcase designed for school students.
•Achieved historical victory by becoming the first team to completely clear the maze
•Facilitated the development of a custom 2-sided custom PCB to accommodate multiple connections in a small space
•Identified bottlenecks and established a streamlined and 35% more efficient calibration and communication pipeline
•Implemented innovative solutions across hardware and software, integrating a Raspberry Pi, Teensy, 9 sensors, and motors into a cohesive unit. Authored and debugged over 2,000 lines of C++ code, ensuring precise control and seamless interaction with the Pacbot environment.
•Developed and troubleshooted a gesture control system using ESPNOW, IMU Sensors and C++, creating an interactive robotics experience for over 25,000 visitors at the Engineering Open House, an engineering showcase designed for school students.
Technotreon Intellectual Ventures: Riot control UAV
Led a team of 4 interns in the development of a cutting-edge riot control drone for enhanced military operations, meeting stringent timelines and budget constraints. By leveraging advanced software tools and implementing patent-focused design methodologies, the team successfully engineered two patent-worthy drone models while significantly reducing the design phase time.
• Utilized KiCad, Fusion 360, 3D printing, and Arduino proficiency to design and build two functional and patent-worthy drone models.
•Streamlined the engineering design process by integrating patent-focused methodologies, resulting in a 30% reduction in design phase time compared to previous projects.
Due to NDA restrictions that is all I can share
• Utilized KiCad, Fusion 360, 3D printing, and Arduino proficiency to design and build two functional and patent-worthy drone models.
•Streamlined the engineering design process by integrating patent-focused methodologies, resulting in a 30% reduction in design phase time compared to previous projects.
Due to NDA restrictions that is all I can share
Autonomous Line Maze Solver
Designed, programmed, and tested a maze-solving robot capable of traversing a line maze and finding the shortest path, featuring a custom sensor array and calibration system for precise navigation.
Optimized PID control algorithms and implemented the right-hand rule along with other logical algorithms to efficiently solve and traverse the maze.
Collaborated with mechanical and software teams to ensure seamless integration of hardware and software components, leading to three national-level competition wins.
Optimized PID control algorithms and implemented the right-hand rule along with other logical algorithms to efficiently solve and traverse the maze.
Collaborated with mechanical and software teams to ensure seamless integration of hardware and software components, leading to three national-level competition wins.
PURE: Undergraduate Machine Learning Researcher
As part of the PURE research program, helped develop a surrogate model for Ion Energy Angle Distributions (IEAD) to replace computationally heavy methods.
• Part of the PURE research program with less than a 1% acceptance rate for my position.
• Worked in the Laboratory of Computational Plasma Physics, under Sonata Valaitis and Mhd. Mustafa, to develop an innovative surrogate model for Ion Energy Angle Distributions(IEAD). This is to replace the current computationally heavy methods.
•Converted thousands of lines of MATLAB code and documentation to Python, achieving a staggering 600% faster training time by leveraging GPU processing. This enhanced efficiency streamlines future research and analysis.
• Optimizing data using algorithms like Coordinate inversion and Polynomial Chaos Expansion to fit it in Sparse Grids of various levels.
• Part of the PURE research program with less than a 1% acceptance rate for my position.
• Worked in the Laboratory of Computational Plasma Physics, under Sonata Valaitis and Mhd. Mustafa, to develop an innovative surrogate model for Ion Energy Angle Distributions(IEAD). This is to replace the current computationally heavy methods.
•Converted thousands of lines of MATLAB code and documentation to Python, achieving a staggering 600% faster training time by leveraging GPU processing. This enhanced efficiency streamlines future research and analysis.
• Optimizing data using algorithms like Coordinate inversion and Polynomial Chaos Expansion to fit it in Sparse Grids of various levels.
Mobiboost- Wheelchair mobility attachment
Engineered a patentable dynamic push-assist attachment for wheelchair users, designed to reduce body ailments caused by unnatural positioning and prolonged use.
Applied patent-focused design principles to create a universal, easy-to-attach system compatible with the majority of wheelchairs, delivering 65% cost savings and 40% weight reduction compared to existing push-assist systems.
Integrated ESP32 and compact brushless motors, making the system highly responsive to user movements, enhancing the overall user experience.
Gained expertise in user research and business solutions by participating in NSF I-Corps and Cozad New Venture Challenge, winning over $1,500 for delivering strong pitches and innovative ideas.
Applied patent-focused design principles to create a universal, easy-to-attach system compatible with the majority of wheelchairs, delivering 65% cost savings and 40% weight reduction compared to existing push-assist systems.
Integrated ESP32 and compact brushless motors, making the system highly responsive to user movements, enhancing the overall user experience.
Gained expertise in user research and business solutions by participating in NSF I-Corps and Cozad New Venture Challenge, winning over $1,500 for delivering strong pitches and innovative ideas.
Combat Robot
Designed and engineered two 30kg war robots for 60lbs Robowars in India with team Xenon, focusing on mechanical design, fabrication, and weapon selection. Youngest member in the team, ever.
•Built a 9kg EN24 drum weapon with a custom gearbox and Al6061 body featuring a wedge-shield structure for optimized performance.
•Enhanced motor cooling with a top plate cutout and secured components using foam and tape for stability in combat.
•Optimized wheel traction for effective operation, even when inverted.
•Won multiple national-level university competitions, showcasing expertise in robotics engineering.
•Built a 9kg EN24 drum weapon with a custom gearbox and Al6061 body featuring a wedge-shield structure for optimized performance.
•Enhanced motor cooling with a top plate cutout and secured components using foam and tape for stability in combat.
•Optimized wheel traction for effective operation, even when inverted.
•Won multiple national-level university competitions, showcasing expertise in robotics engineering.
Concept UAV Design
Examined existing designs, designed and proposed three different autonomous UAV designs over almost 3 years for package/vaccine delivery. Won 6+ national and international competitions and gained media coverage as well.
Worked on:
1. Electronics: Researched and finalized between various ways and specific components for multiple purposes:
•Obstacle avoidance using a combination of photogrammetry(cameras), radars, and lidars, with the specific components, computational power, and overall coverage in mind.
•Calculated the power system configuration considering the lift, maximum speed, and power reserve into consideration.
•Selected components for redundant communication and location tracking of the UAV. Considering aspects like loss of signal, the urban canyon effect, etc.
•Designed the final circuit diagram in Fritzing.
2. Simulations: Did Computational Fluid Dynamics(CFD) simulations using Fusion 360 and Simscale to calculate and analyze results like the lift to drag ratio, coefficient of lift, and coefficient of drag.
3. Business Case: Examined the real-world business feasibility of deliveries using autonomous UAVs. Presented business cases based on the specific challenge requirements.
4. Design: Researched various designs, compared them using efficiency, speed, challenge feasibility, etc., and selected designs based on the challenge.
Important: Media for only one project is share as all of them are similar in nature
Worked on:
1. Electronics: Researched and finalized between various ways and specific components for multiple purposes:
•Obstacle avoidance using a combination of photogrammetry(cameras), radars, and lidars, with the specific components, computational power, and overall coverage in mind.
•Calculated the power system configuration considering the lift, maximum speed, and power reserve into consideration.
•Selected components for redundant communication and location tracking of the UAV. Considering aspects like loss of signal, the urban canyon effect, etc.
•Designed the final circuit diagram in Fritzing.
2. Simulations: Did Computational Fluid Dynamics(CFD) simulations using Fusion 360 and Simscale to calculate and analyze results like the lift to drag ratio, coefficient of lift, and coefficient of drag.
3. Business Case: Examined the real-world business feasibility of deliveries using autonomous UAVs. Presented business cases based on the specific challenge requirements.
4. Design: Researched various designs, compared them using efficiency, speed, challenge feasibility, etc., and selected designs based on the challenge.
Important: Media for only one project is share as all of them are similar in nature
JackSafe: Safety Jacket
Engineered and prototyped a women’s safety jacket, integrating a shock module with a GPS-GSM based alarm and live messaging system for real-time location tracking and emergency alerts.
Designed the circuit using an integrated Boost Converter and GPS-GSM modules to enhance safety features, ensuring reliable and efficient functionality.
Published a research paper titled "Electronic clothing development using Boost Converter Circuit and GPS-GSM Module with Arduino," showcasing the innovative integration of electronics in wearable technology.
Designed the circuit using an integrated Boost Converter and GPS-GSM modules to enhance safety features, ensuring reliable and efficient functionality.
Published a research paper titled "Electronic clothing development using Boost Converter Circuit and GPS-GSM Module with Arduino," showcasing the innovative integration of electronics in wearable technology.
BikeSafe: Mottier Innovation Challenge
Conceptualized and prototyped BikeSafe, a revolutionary self-installable bike anti-theft device, within a
limited $500 budget. This innovative solution leverages a three-tiered security system:
•Electric shock deterrent: Discourages would-be thieves with a controlled shock mechanism.
•GPS tracking: Provides real-time location monitoring for stolen bikes.
•Integrated alarm and smart app: Alerts owners and broadcasts location upon unauthorized movement.
Led market research and competitor analysis, identifying key consumer trends and target markets. This data-driven approach ensured BikeSafe addressed critical user needs and stood out from existing anti-theft options.
Championed resourcefulness and cost-effectiveness. Engineered a highly compact PCB with 70% reduced footprint, maximizing functionality within budget constraints. Utilized 300+ lines of C++ code and conducted multiple user tests to refine the prototype.
Demonstrated exceptional teamwork and leadership as part of an all-freshman team. Successfully navigated the competition, earning an honorable mention and $500 prize – a testament to BikeSafe's potential and the team's dedication.
limited $500 budget. This innovative solution leverages a three-tiered security system:
•Electric shock deterrent: Discourages would-be thieves with a controlled shock mechanism.
•GPS tracking: Provides real-time location monitoring for stolen bikes.
•Integrated alarm and smart app: Alerts owners and broadcasts location upon unauthorized movement.
Led market research and competitor analysis, identifying key consumer trends and target markets. This data-driven approach ensured BikeSafe addressed critical user needs and stood out from existing anti-theft options.
Championed resourcefulness and cost-effectiveness. Engineered a highly compact PCB with 70% reduced footprint, maximizing functionality within budget constraints. Utilized 300+ lines of C++ code and conducted multiple user tests to refine the prototype.
Demonstrated exceptional teamwork and leadership as part of an all-freshman team. Successfully navigated the competition, earning an honorable mention and $500 prize – a testament to BikeSafe's potential and the team's dedication.
bottom of page