Kitchen Chaos (Clone)

Kitchen Chaos is a Unity-based game project inspired by a course from the CodeMonkey YouTube channel. This project follows a step-by-step approach to building a game from scratch while covering a range of key game development concepts, including:

• Input System
• Post Processing
• User Interface Design
• Scene Management
• C# Scripting
• Character Movement & Interactions
• Animations
• Shader Graphs

Check out a demo video of my project on YouTube! 👇

eZeGo!

eZeGo is an interactive, real-time stage lighting software that enables intuitive music visualization. It is designed for use in both physical venues, from small rooms to large concerts, and virtual environments, including traditional 3D games and immersive VR/XR spaces.

eZeGo is built from the ground up with performance and responsiveness in mind. It supports all major platforms (Windows, Linux, macOS) and is free to use for personal and commercial use under the MIT license.

Qosmit2D

Qosmit2D is a simple 2D game engine written in C++ and SFML. It uses Entity-Component-System for managing game objects. Currently Qosmit2D has implemented bellow systems:

• ECS : to handle game objects (entities) within the game and their behavior (components) through centralized logic blocks (systems)
• Resource Manager : for loading and managing assets (images, sprites, audio files, etc) and scenes
• Input System : to process mouse and keyboard events and map them into Commands
• Audio Player : to play music and sound effects
• Collision Physics : to handle collision for simple 2D objects
• Animation : to drive simple transform and sprite animations
• Visual Debugger : to provide insights from various game engine system while the game is running
• Frame Profiler : used Tracy to profile runtime and memory usage of various parts of the engine

I created this game engine as a learning experiment to explore various concepts discussed in the Game Engine Architecture book and understand how different engine systems interact. It is not intended for production use at any scale.

Obsidian SVG Styler

SVG Styler is a simple Obsidian community plugin designed to help you easily modify the style properties of SVG vector graphics directly within the editor. With this plugin, you can customize properties like color, border, opacity, and more. Supported drawable tags include: path, line, rect, circle, ellipse, and polygon.

Traffic Accidents Visualizer

Traffic Accidents Visualizer is an interactive web application that helps users explore traffic accident data in British Columbia by municipality, year, and contributing factors. By combining dynamic maps and rich Sankey diagrams, this tool offers an intuitive and visual way to understand how accident causes vary across locations and time. The application is powered by a FastAPI backend that reads and processes publicly available traffic accident datasets on demand. The frontend is built using jQuery, Google Charts, and Leaflet.js to deliver a responsive and interactive user experience. For geographic visualization, we overlaid municipality boundaries onto the Leaflet map using custom GeoJSON and shapefiles of British Columbia. These files were processed and cleaned using GeoPandas, including a projection conversion from EPSG:3857 (Pseudo-Mercator) to EPSG:4326 to ensure proper alignment with web map standards

LightKnight

LightKnight is a desktop-based LED strip animator developed using web technologies. It enables users to configure multiple LED channels (either RGB or single color), import a music file, and design animations synchronized with the audio.

Check out a demo video on YouTube (apologies for the low recording quality 🤷‍♂️).

This project was archived in October 2020 and is no longer maintained. I am currently working on eZeGo, the successor to LightKnight.

Rice Seed Classifier

For my Computer Vision Engineer Internship at Rice Research Institute of Iran (RRII), I worked on automating rice seed classification and defect detection. The goal was to identify different local rice varieties and spot broken or defective grains faster and more accurately than human experts. Using deep learning and computer vision, I trained CNN models that achieved 87.1% accuracy in variety classification and 97.6% accuracy in defect detection while making the process 120x faster. I used OpenCV with the Canny algorithm to automatically isolate and crop individual rice grains, then built a dataset of over 1 million labeled images. The models were trained and deployed with TensorFlow and Keras on an Ubuntu environment.

A big part of this project was dataset creation and automating the process of finding and cropping single rice grains while filtering out empty slots and broken seeds. The full dataset and source code are still confidential, but I’ve shared a simplified version for educational purposes.

Fear of God

This project showcases a visually stunning mobile application UI/UX design, originally created by Lisa Parker on Dribbble. The design incorporates dynamic elements like Hero animations and a combination of vertical and horizontal list views. Built using Flutter, this demo leverages its powerful UI, typography, and animation capabilities to bring an engaging and seamless user experience to life.

To see this demo in action, check out this GIF in my project's GitHub repository.

IoT Humidity Slider

This demo project is inspired by an original design from Elijah Bozniak on Dribbble. It showcases a mobile application for controlling home humidity, featuring sleek user interface and fluid animations.

I selected this design to highlight UI animations and interactions within React Native using the Reanimated library. Reanimated is a robust animation library that enables developers to create intricate, interactive, and high-performance animations directly on the UI thread.

To see the IoT Humidity Slider app in action, check out this GIF in my project's GitHub repository.

Genetic Programming Solver

This project explored a specific subset of heuristic algorithms known as Genetic Algorithms, which are inspired by the process of natural selection. Genetic Programming, a related technique, evolves computer programs or mathematical expressions by iteratively selecting, mutating, and recombining solutions to optimize performance based on a set of defined fitness functions.

My goal was to apply these techniques to solve problems with a tree-structured nature, such as circuits, mathematical expressions, and decision trees. This solver takes an input-output mapping and generates a function—either discrete or continuous—that best fits the given data, effectively acting as a curve-fitting algorithm optimized for hierarchical expressions.

For my Digital Circuits course project, I utilized this solution to generate custom circuit components, significantly enhancing my design process. This project was recognized as the best in the class.

Optimal Triangulation

This project was developed for my Advanced Algorithms course, where my team and I designed and implemented an algorithm to triangulate a convex polygon in a way that minimizes the total length of internal edges (i.e., the sum of triangle perimeters). We achieved this optimization using a dynamic programming approach.

Beyond algorithm design and implementation, I also developed a GUI desktop application to visually demonstrate our optimizer in action, allowing for an interactive and live presentation in class.

ARGOSTA.ME