Hi, I'm Semih, a frontend-focused developer from Turkey with a strong interest in machine learning. I enjoy building interactive systems, performance-oriented applications, and AI-powered products.
DJ-Controller: Building a digital DJ controller experience that gets as close as possible to real-world hardware interaction.
Cithus: An everyday desktop assistant with voice control. Powered by a Node.js backend and a Rust (Tauri) frontend for high performance.
Expert Dashboard: A management dashboard for mental health professionals to manage reservations and client interactions.
Pose (Coming Soon): A React Native-based social media platform focused on teaching posing techniques.
Cs2nades.me: A CS:GO forum dedicated to grenade lineups and tactical sharing.
WeddingShare: An event-based media collection platform allowing guests to upload photos and videos.
Locpik: A location-based social media platform.
WebAmI: A VR-like interactive world built in JavaScript using MediaPipe with client-side AI processing.
document-workspace:A browser-based document workspace for uploading, organizing, and querying PDF documents using AI. Built with React and TypeScript.
Just Note: A simple Chrome extension that lets you take notes per URL. Each page gets its own notes.
deva:DevAlibi - A global CLI tool to quickly launch and manage your local development projects with a single command.
Minesweeper: Classic Minesweeper game implemented with JavaScript.
Seesaw: A physics-based browser game built with JavaScript.
Chatbot Interface: A clean, minimal, and reusable chatbot UI design.
Movies: A Netflix-inspired movie browsing interface.
WeatherApp: A simple and intuitive weather application.
Neural-Network-From-Scratch: A fully custom Artificial Neural Network implementation in Python supporting forward and backward propagation without external ML libraries.
Sentiment Analyzer: A sentiment analysis system built on top of my custom neural network implementation.
MoRe: A hybrid recommendation system combining content-based and personalized filtering for tailored movie suggestions.
R2TD: A real-time traffic analysis system using YOLOv8 and OpenCV for vehicle detection, tracking, and lane-crossing monitoring.