Haamed Rahman Email & Phone Number

Amherst, Massachusetts, United States

As the Lead Machine Learning Engineer on the UMass Mechatronics Team for the 2024-2025 ASME Student Design Competition, I am responsible for developing an autonomous algorithm for sorting, identifying, and packaging ball bearings of varying sizes and materials. Leveraging XGBoost, I have currently achieved a 94% accuracy rate, with further optimization underway. The GitHub repository for this project remains private due to the ongoing nature of the competition.In this project, which tackles the challenges of modern assembly lines where speed, precision, and efficiency are paramount, my primary role is to design and implement a machine learning-based system that reduces the sorting device's overall size and complexity by 60% while enhancing processing speed by 200% by eliminating delays. The algorithm is capable of automatically identifying different materials and sizes, directing them to six designated packaging stations.Working alongside a multidisciplinary team of mechanical, computer, and electrical engineers, I am leading the integration of this software into embedded systems, ensuring seamless, real-world functionality. This project focuses on delivering a compact, reliable solution that meets high-throughput requirements, enhances assembly line efficiency, and lowers operational costs, ultimately addressing the need for cost-effective, precision-driven automation.

As a Software Engineering Intern at Workable Space, I researched and implemented pod automation features using Flask and Spring Boot in a virtualized Linux environment. I was responsible for managing pod control systems with RESTful APIs, enhancing the remote control functionality, and improving system efficiency. Additionally, I utilized Linux VMware, Docker, and Git to streamline testing and deployment processes, ensuring seamless integration and faster development cycles for pod automation solutions.

As a Research Assistant under Professor Ken Khan, I developed a convolutional neural network (CNN) model for fire detection, achieving 85% accuracy with real-time image upload capabilities using Tkinter. My work focused on enhancing safety protocols through computer vision techniques, including data augmentation and reducing model overfitting by 12%. I also implemented advanced image preprocessing techniques to improve the model’s robustness under varying lighting conditions, ensuring more reliable fire detection in dynamic environments.

Bachelor Of Science - Bs, Computer Science

High School Diploma & Ap Capstone Diploma, 9-12

Activities and Societies: Python Club (Co-President) | Science Olympiad (Co-Founder and Vice President) | Developers Club(Executive.