Text To Shape AWS Project

This project combines cloud engineering, AI, and security on AWS. Primarily for fun and learning. I built a Flask API on EC2 that takes natural language input, uses AWS Bedrock to turn it into a JSON command, and then draws the requested shape with Python Pillow. The image is saved in S3, and users get a presigned URL that expires after 10 minutes. To keep it secure, I applied IAM least privilege, input validation, rate limiting, and CloudWatch monitoring. I could have used services like Lambda, SageMaker, or Route 53 with a domain and HTTPS, but I avoided them to stay simple and within Free Tier.

AWS Cloud Engineering AI Engineering StreamLit Python S3 Bucket EC2 Server Bedrock AI IAM
Match3 Screenshot 1 Match3 Screenshot 2 Match3 Screenshot 3 Match3 Screenshot 4 Match3 Screenshot 5 Match3 Screenshot 6 Match3 Screenshot 5 Match3 Screenshot 6

Match3 Game

This Unity project is a match3 game which has 10 levels. Levels are designed with JSON files. It has a main menu, a level selection screen and a game over screen. Project uses Scriptable Objects, event system, addressables and object pooling for better optimization. It has cubes, rockets, obstacles, particle effects and sprites. It correctly manages to wait until all animations finish on the scene.

Unity C# Game Development Mobile Game Game Design 2D Animation
Match3 Screenshot 1 Match3 Screenshot 2 Match3 Screenshot 3 Match3 Screenshot 4 Match3 Screenshot 5 Match3 Screenshot 6

Unity VR Game Engine

This project is a VR game engine made with Unity. It is a block coding game that allows users to create their own games. It provides an easily understandable environment in which users can put assets from its library, manipulate assets' properties and create their own game. They can share and visit other users' games. Or they can share the code to open their game in VR mode.

Unity C# Game Development VR Block Coding Google Cardboard
Project Screenshot 1 Project Screenshot 3 Project Screenshot 2 Project Screenshot 4

Find What To Do - WhatNow

This is a mobile application in IOS Store made with bare react native. Application uses device id to keep track of users' data. At first start, application asks 16 questions to know the user's personality. Then user can open the app and choose between what to eat or what to eat. Application pulls data from weather, location and ticketmaster api and creates a prompt for ChatGPT API. They get 3 recommendations according to their personality with weather and location. Like dating applications, user can like or dislike the recommendations to match with them.

IOS RevenueCat React Native Supabase AI ChatGPT API Location API Weather API Ticketmaster API
Project Screenshot 1 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4

Banana Photo Agent - BananAgent

This is a mobile application in IOS Store made with Swift. When Gemini has announced their new AI "Nano Banana", I noticed it is quite powerful and decided to use its API to make a mobile application. Users can upload images and enter prompt to generate new image using API. They can see their own generated images and they have a discovery screen. For each generation, users spend their credits. They can buy credits via paywall.

IOS RevenueCat Swift Supabase AI Gemini API
Project Screenshot 1 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4 Project Screenshot 4

CBIR for fashion

This project uses huge database of fashion images to create a CBIR (Content Based Image Retrieval) system. It lets user to choose an image (product) and finds similar images from the database. It uses classic methods and CNN model to extract features from images and a KNN model to retrieve similar images. It also uses ML Tuning to find the best hyperparameter for the model. It uses several accuracy metrics to measure the success rate. It uses labels of found images to measure the accuracy. It also labels the chosen image.

Python TensorFlow Image Processing CBIR ML Tuning Accuracy Metrics
Project Screenshot 1 Project Screenshot 2 Project Screenshot 3

Q1 Journal Paper

This master course's term project is about conducting experiments on students using Meta Quest 3 for daily tasks. Several tasks were defined, and students were asked to perform them with and without Meta Quest 3. The collected data was analyzed, and the results were presented in an article paper supported by more than 20 references. The article has been submitted to the International Journal of Human-Computer Interaction for publication. The journal is ranked as Q1, indicating its high impact and recognition in the field of Human-Computer Interaction research.

VR/AR Meta Quest 3D Development Q1 Journey Article Data Analyses Journal Research
Project Screenshot 1 Project Screenshot 2 Project Screenshot 3

Codeway Website

This project is a basic mockup of codeway web. It is a full-stack configuration management system that allows application managers to log in through a responsive web panel. Users can edit app configuration parameters stored in a Firestore database, and serve these configurations via a secure REST API to mobile clients. The panel, built with Vue 3 and Firebase Authentication, prevents concurrent edit conflicts, send updates to the backend with Firebase ID tokens.

Vue Node.js Vercel Heroku Cloud Firebase
Project Screenshot 4 Project Screenshot 4

Whispurr - Dating App

This is an mobile app made with Flutter. Users signup via email, choose their spirit animal and create their profile. Then in discover page they can find other users filtered by preferences. They can WHisper (message) to other users. If they like each other, they can match and start a chat. If they reach to 10 mutual message limit the photos become visible.

Flutter Dart Firebase
Project Screenshot 6 Project Screenshot 7 Project Screenshot 8 Project Screenshot 9 Project Screenshot 1 Project Screenshot 2 Project Screenshot 3 Project Screenshot 4 Project Screenshot 5

One Minute - Daily Task App

This is an mobile app made with Flutter. Users signup via email, setup their profile by answering onboarding questions. Then everyday at a specific time, all users get notification about they get a daily task. Tasks can be completed in around 1 minute. The aim is to make people less stressful and help their phone addiction. They level up and earn badges as they complete the tasks. They can add friends and see each others' profiles.

Flutter Dart Firebase
Project Screenshot 1 Project Screenshot 2 Project Screenshot 3 Project Screenshot 5 Project Screenshot 6 Project Screenshot 7

Gesture Detection

This project is a gesture detection system made with OpenCV and MediaPipe. It lets users to use their hand gestures to control the computer. They can open the gallery, scroll 4 directions, select and click on an image. They can zoom in-out or rotate the images via face camera.

Computer Vision Machine Learning Python OpenCV MediaPipe Gesture Recognition
Project Screenshot 1 Project Screenshot 2 Project Screenshot 3

File Management Tools

A comprehensive desktop application for mac and windows that made and put on sale in Grumroad. It provides powerful file management capabilities including batch file renaming, PDF manipulation, and image processing. The app allows users to rename dozens or hundreds of files in seconds with custom base names, preview results before committing, and undo mistakes. For PDFs, users can split or extract specific pages, merge multiple PDFs, delete unwanted pages, and handle password-protected documents. The image processing feature enables quick resizing with custom dimensions, format conversion, and preview functionality. All operations are performed offline, ensuring user privacy and data security.

File Management Desktop Application PDF Manipulation Image Processing Offline Processing
File Management Tools Screenshot 1 File Management Tools Screenshot 2 File Management Tools Screenshot 3 File Management Tools Screenshot 4

AI Templates Bundle

A comprehensive digital product bundle featuring six AI-powered productivity tools and templates that made and put on sale in Grumroad. The collection includes a 1-month goal planner for structured achievement, an ultimate email template toolkit with 50+ professional templates, a YouTube video ideas vault for content creators, an AI income guide for passive revenue generation, an AI interview coach with ChatGPT prompts for job seekers, and 50+ AI prompts specifically designed for content creators across multiple platforms. Each product is designed to leverage AI technology to enhance productivity, creativity, and professional development.

AI Integration Productivity Tools Content Creation Digital Products ChatGPT Prompts Template Design
AI Templates - Goal Planner AI Templates - Email Templates AI Templates - Content Creation AI Templates - Interview Coach AI Templates - AI Prompts AI Templates - Screenshot 2