A flagship initiative by the Khoury Robotics Club, WaltZ is a 6-DoF bipedal robot driven by an RP2040 microcontroller and 9-axis IMU. Built to be affordable and easy to fabricate, this ongoing project will serve as an educational tool in robotics courses, showcasing core principles of gait, balance, and embedded control.

Used a custom genetic algorithm to train a simulated robotic arm to evade flying objects. Developed solely in NumPy and Pygame without external libraries, the project introduces a multi-stage scoring function that incrementally increases task difficulty, substantially speeding up training compared to traditional approaches.

A web-based platform for real-time tracking of MBTA-operated buses. Offers intuitive route-based visualization and dynamic location updates, providing valuable insights for transit management and daily commuting.

A final-year undergraduate project aimed at reconstructing 3D point clouds from 2D camera images. Explored both stereoscopic methods and machine learning approaches, furthering research in image processing and computer vision.

Designed and 3D-printed a small-scale model of NASA’s Mars Helicopter featuring counter-rotating propellers. While demonstrating rotor feasibility, the prototype highlighted sensitivity to the craft’s center of gravity (CG), underscoring key engineering challenges in rotary-wing UAVs.

A hackathon project that leverages generative adversarial networks (GANs) to interpolate between existing frames, effectively doubling the frame rate of a source video. Demonstrates potential applications in bandwidth optimization and responsive video streaming.

Developed a trie-based inverted index to efficiently search large-scale text data. Compared the approach against Elastic Search in terms of accuracy and performance, offering insights into scalable data-structure design for information retrieval.

A 2D particle simulator illustrating the Kinetic Theory of Gases. Explores various collision-detection and resolution algorithms, visualizing core concepts in physics and computational modeling.

Implemented a generic framework for neural network training using genetic algorithms and simulated annealing. Focused on classification tasks, demonstrating the potential of evolutionary computing approaches in AI.

Built a MongoDB-based Database as a Service (DBaaS) platform featuring auto-scaling and fault tolerance. Designed for dynamic resource management, ensuring high availability and performance in data-centric applications.