Projects

A WiFi CSI-based fall detection pipeline built with Sam Desai and Kristine McLaughlin at the University of Michigan, running spectrogram classification on a Raspberry Pi.

A hierarchical caching system built with Arnav Reddy, Muzhe Wu, and Yuxuan Liu, at the University of Michigan that speeds up multimodal LLM inference by reusing responses and intermediate state across visually or semantically similar requests.

A study built with Francisco Haas, Siddharth Kalra, and GaHyun Yoon at the University of Michigan on how textual adversarial attacks fool spam classifiers, and how adversarial training defends against them.

I created this webpage from scratch, using many different languages to do so. The server runs on AWS Lambda functions written in Go.

My implementation of the Elementary Cellular Automaton, created in Rust with the Bevy engine and compiled to Web Assembly.

My implementation of Craig Reynolds' Boids. It was created with Rust, using the Bevy game engine, and was compiled to Web Assembly. This is my first Rust project

I researched how a globally optimized course schedule could increase student satisfaction. Supervised by Professor Amir Ali Ahmadi (Princeton).

This is my recreation of the Google Chrome Dinosaur Game. It is created with the pygame library and includes sound.

A simple sudoku solver and generator. The solver is created in Python and solves using a recursive, depth-first search. The graphics are created using the Python library: pygame.

My version of TicTacToe, created in Python using the Python library: pygame.

To practice using Java, I recreated Google's version of the classic game Snake. I used the AWT and Swing libraries to accomplish this.

My very first project: I created Pong using Java. I used Java's AWT and Swing libraries for the graphics.