I'm Fied Elahresh

Developing an AI interview web app that enables users to simulate real-time mock interviews with AI, featuring WebSocket-based interaction, JWT authentication, and secure password hashing with bcrypt.

Designed a real-time multiplayer 3D maze game using Java sockets and Java3D, implementing synchronized gameplay, collision detection, NPC behavior, and camera tracking for smooth client-server interaction.

Built a machine learning pipeline using Python and PostgreSQL to predict deck matchups across 200k+ matches, improving accuracy by 15% with antisymmetric features and GPU-accelerated training.

Created an AI-powered Bitcoin trading bot that combines sentiment and technical analysis (RSI, SMA) to make trading decisions, achieving $30,000 in backtested profit.

End-to-end ML project predicting house prices from tabular features, including data cleaning, feature engineering (encoding, scaling), model training and cross-validation, and a reproducible sklearn pipeline.

Developed a Duolingo-style app that teaches foundational Islamic knowledge, featuring A/B tested learning modules, a Supabase backend, and integrated RevenueCat for in-app subscriptions.

Released a subscription-based Islamic mobile app with Quran reading, daily hadiths, and an AI-powered chat assistant using Supabase and OpenAI API, available on the App Store.