Projects
Inferno Tower - Stroboscopic Effect Fountain
Feb 2026 - Ongoing
Overview
This project is a lighting design project inspired by the Inferno Tower from Clash of Clans, designed for the Lutron Lighting Competition. It recreates the stroboscopic effect, where precisely timed pulsing lighting makes flowing water appear levitating or moving in slow motion. The goal was to create an interactive display that combines visual aethetics, engineering principles, and embedded systems.
Process
The Inferno Tower was 3D modeled in Blender then 3D printed to house all components cleanly. We used an Arduino Nano to control the strobing lights and a PWM-driven water pump. A rotary encoder, mini OLED display, and push button were used as UI to intuitively control timing parameters like pulse frequency and pump power. Components were initially soldered to a perfboard, but we plan on using a custom PCB for a more finalized design.
Results
The project is still ongoing. We plan on improving our prototype despite the competition ending. We strengthened our skills in rapid prototyping, system integration, and adapting to critical issues like part delays.
Portfolio Website
Jan - Apr 2026
Overview
An interactive 3D portfolio for displaying past and ongoing projects in an aesthetic scene. The scene is my (ideal) room modeled in Blender, rendered with Three.js and WebGL, features clickable objects, animated elements, and light/dark modes. My idea for this project was to create a functional portal to showcase my work, while keeping the website fun and engaging.
Process
The room was modeled and baked in Blender with image texture sets to avoid real-time lighting costs. I used WebGL to handle scene rendering, Three.js for texture and glTF model loading, and OrbitControls for camera movement. Raycasting and GSAP makes the website more interesting through its animations: modal transitions, object transformations, and a custom intro scene. GPU preloading day/night textures creates seamless transitions between light/dark modes.
Results
Building this was one of the most fun and valuable projects I've worked on: went from complete beginner in Blender and Three.js to having a polished 3D website. Designing my portfolio taught me a lot about Blender, Three.js, GSAP, and web development. The website UI feels intuitive, smooth, and finalized. I'm continuing to develop new features and improvements for the website.
Social Network Graph App
Nov - Dec 2025
Overview
C++ desktop application built for EC327 (Intro to Software Engineering) that simulates and visualizes a social network of students. Users can load or generate datasets, explore social adjacency, and find minimum degrees of separation between any two people. The project was developed by a team of five using Agile practices, with a Qt-based GUI layered on top of a modular, dependency-injected backend.
Process
Project workload was split up into small tasks and organized in a Kanban board. Team members assigned themselves to tasks and helped others when finished. The core simulation in SocialGraph.cpp treats each Person as a node and builds directed edges based on a configurable similarity score across attributes like region, graduation year, favorite colors, and engineering focus. Users adjust category weights and a top-N neighbors setting to control which edges are kept. Multiple PersonReader implementations (CSV, HTTP+JSON, and random generation) are wired in via concepts learned in class (Inversion of Control and Dependency Injection). A Qt interface connects this logic to a graphical view, letting users load, edit, persist, and explore datasets while running BFS-based pathfinding and traversal under the hood.
Results
Delivered a fully interactive social graph application with directed, weighted edges driven by a configurable similarity model. The system supports CSV, HTTP+JSON, and synthetic data sources, along with autosave, reload, and in-place editing of Person records. Users can visualize the network in real time, navigate by clicking nodes or entering IDs, and highlight shortest paths between arbitrary people. The final application met all EC327 requirements, including modular data loading, persistence, configurable edge logic, and a documented simulation-generated social graph.
Truss Simulation & Optimization
Nov - Dec 2025
Overview
EK301 Engineering Mechanics 1's final project was a truss design and simulation. The goal was to design a truss structure that maximizes load capacity while meeting material, geometric, and cost constraints. Our primary design philosophy was to enhance a preliminary design that met all the design constraints and optimizing it the best we could. Optimizing meant redistributing forces within the truss, which required simulation and modeling of internal forces. MATLAB was utilized to rapidly analyze different design iterations, allowing for us to create a strong, valid truss design.
Process
Our design process began with a preliminary truss that satisfied all constraints. Using our MATLAB analysis program, we calculated internal forces, member buckling strengths, and the maximum load each design could support. I created another MATLAB script to edit truss geometries in real-time, which further sped up design iterations, bypassing the manual computations of joint positions and verifications of member lengths. After creating multiple designs significantly above the minimum requirement, I tried to optimize it even further. I wrote another MATLAB script for a naive Monte Carlo simulation to shift each non-fixed joint position, and repeatedly ran the script until the max capacity peaked.
Results
The final optimized truss design achieved a maximum load of 57.8 oz, with an uncertainty range from 53.1 to 59.8 oz. Forces in all members were kept below their respective buckling limits at the target load. The total cost of the truss was underbudget at $188, resulting in a load-to-cost ratio of 0.307 oz per dollar. This project demonstrated how MATLAB can be used as a powerful tool for structural analysis and optimization. Our interactive editor further allowed us to visualize and manipulate truss configurations, streamlining the design and testing process. The methodology and results highlight a practical approach to maximizing structural performance within given constraints.
Automatic Liquid Dispenser
Sep - Dec 2025
Overview
EK210 Hands on Engineering Final Project: a contact-free automatic liquid dispenser designed for home, medical, and public settings to promote hygiene and convenience. The device uses an infrared sensor to detect user presence and triggers a micro water pump controlled by an Arduino to dispense precise amounts of liquid. An OLED and 16x2 LCD display provide real-time feedback on system status and remaining liquid, making the device intuitive and user-friendly.
Process
The enclosure was designed in CAD and fabricated using 3D printing and acrylic panels to securely house all electronics and the liquid reservoir. A 5V IR distance sensor detects objects within a 3-11 cm range, and the Arduino triggers a 12V Uxcell micro water pump via a MOSFET to dispense liquid. Firmware handles timing, pump control, and display updates, while a buck converter ensures safe power distribution to all components. Componenets were carefully wired and soldered to ensure reliability and compact packaging.
Results
The prototype dispenses liquid in approximately 1.62 seconds with over 95% reliability and can handle different viscoscities (water, soap mixtures) allowing for adaptable use. The device is compact (6 x 7.3 x 13 in), lightweight (1.24 kg), and easy to refill, clean, and operate. Testing confirmed accurate dispensing, consistent sensor detection, and under budget at $106.30, achieving our client's goals. This project reinforced skills in embedded systems, sensor integration, circuits fabrication, and user-centered design.
Temperature Monitoring Device
Apr - May 2025
Overview
First engineering project at BU for EK131 Hands On Engineering. An Arduino-based temperature monitoring system designed to measure ambient temperature and alert the user when readings fall outside a configurable range. The device combines embedded software, circuits, and mechanical design in a device. Provides both a visual and audible notification in a compact, portable enclosure.
Process
The enclosure was designed in Onshape and fabricated via FDM 3D printing to securely house all components, including the Arduino Uno, TMP36 sensor, LEDs, buzzer, and a 9V battery. Firmware written in Arduino C reads the analog output from the temperature sensor, converts it to Celsius and Fahrenheit, and displays the readings on an LCD. If the temperature exceeds user-set thresholds, the code triggers a red LED and buzzer for alerts. Internal wiring was assembled and soldered for reliable connections, with power supplied by a 9V battery or external USB connection.
Results
The completed system successfully measures temperature, displays real-time readings, and alerts the user when values fall outside the set range. The device operates reliably for over eight hours on battery power, with the threshold easily adjustable in the firmware. This project reinforced practical skills in embedded systems programming, electronics assembly, and CAD-based fabrication, providing a strong foundation for future hardware and mechatronics projects.