Projects

A collection of my projects showcasing my work in machine learning and computer graphics.

LoLytics

A full-stack analytics platform built from the ground up to provide actionable League of Legends insights. It combines end-to-end machine learning pipelines, from data collection and model training to real-time inference and visualization.

PyTorch, React, Vite, Seaborn

View on GitHub

Drast

A tool that can rasterize vector graphics onto a 2D canvas in a way that preserves differentiability and allows gradient computation.

Rust

View on GitHub

Color Classifier

A neural network that classifies colors based on human perception, trained on real user data collected from an online giveaway.

JavaScript, TensorFlow

View on GitHub

Super Resolution GAN

Implementation and improvement of Nvidia's SRGAN architecture for image upscaling. Focuses on perceptual loss and generative consistency for sharper, more realistic results.

Python, PyTorch

View on GitHub

Procedural Planets

Real-time procedural planet generation in the browser using WebGL 2.0. Implements noise-based terrain, atmosphere rendering, and adjustable parameters for terrain shape and climate simulation.

WebGL, JavaScript, GLSL

View on GitHub

AI Denoise

In this project, I try several techniques to train a UNet to remove noise from images that have been generated using a raytracing pipeline with number of samples per pixel.

C++, OpenGL, PyTorch

View on GitHub

Meadow

A desktop application to be used as a voxel editor to construct 3D works of art using arbitrarily colored blocks. The original intent behind this project was for it to be used alongside a generative adversarial network to experiment with 3 dimensional output.

C++, OpenGL, GLM, GLFW

View on GitHub

TheTowers

Open source multiplayer minecraft-like team-based game much like capture the flag. Incorporates many of the mechanics present in minecraft such as block placing/breaking, crafting, climbing etc but also adds more such as shooting and sliding.

C++, OpenGL, GLM, SFML, STB

View on GitHub

Melon

A fast, simple, and flexible C++ library that combines the idea of genetic algorithms with neural networks while supporting many common selection, mutation and crossover mechanisms.

C++

View on GitHub

NeuroBird

Using Neuro-evolution to teach artificial neural networks how to play flappy bird.

Javascript, TensorFlow

View on GitHub

In-Painting

A generative adversarial network that learns to regenerate arbitrarily masked sections of an image.

Rust

View on GitHub

WebChess

Web platform for improving chess skills through puzzles, endgames, and opening practice. Integrates a Stockfish backend for analysis and interactive training.

JavaScript, Python

View on GitHub

Small AI Projects

Developed a collection of AI projects as part of a machine learning bootcamp, including an image colorization GAN, DeepDream in PyTorch, text generation with GRUs, sentiment analysis with LSTMs, transfer learning using pre-trained CNNs, and language translation with seq2seq models.

Python, PyTorch, Pandas, HuggingFace

View on GitHub

Discord Filters

An open source face mask/filter for discord video chat. Combines a face detection model with state of the art image to image translation to add custom filters to videos in real time.

Python, PyTorch, OpenCV

View on GitHub

Raytracer

An Offline GPU Accelerated Raytracer in C++ that supports multiple backends. The API for the rendering pipeline was heavily inspired from the raytracing series by Peter Shirley.

C++, Vulkan, OpenGL

View on GitHub