Neural Network Chess Bot | TensorFlow, Keras, Python, Flask, JavaScript • Developed a deep learning-based chess engine capable of evaluating and predicting optimal moves from baseline board positions. • Parsed large-scale, unstructured chess datasets in PGN format, converting static configurations into multi-dimensional 8 × 8 × 12 tensor representations. • Trained a deep Convolutional Neural Network (CNN) via TensorFlow/Keras, yielding a top-1 move prediction accuracy of 25.9%. • Exposed engine logical functions via a Flask API, integrating the backend with an interactive, real-time client-side JavaScript UI. [GitHub] Real-Time Chat System | FastAPI, WebSockets, MongoDB, JavaScript • Architected a scalable, asynchronous chat application utilizing WebSockets and FastAPI to enable low-latency message streaming. • Modeled a MongoDB layer for message persistence, enabling rapid document queries and horizontal state scaling across user sessions. • Designed and implemented full-duplex user features including live network presence tracking, real-time typing indicators, and cross-device session syncing. [GitHub]