My Projects

A fully object-oriented music streaming platform designed to simulate a real-world service like Spotify or Apple Music. It features dynamic user interaction, personalized recommendations via collaborative filtering, artist and album management, playlist creation, and royalty calculations. The system tracks listening behavior, supports premium subscriptions, and includes advanced features like mood analysis, discovery scoring, and year-end user statistics — all orchestrated through a modular class-based architecture.

A modular object-oriented social network simulation featuring user profiles, posts, comments, and engagement analytics. Users can follow, block, and interact with each other through text, image, and video posts. The system supports private profiles, mutual follower detection, trending content, hashtag analysis, and nested comment threads. Includes platform-wide features like user registration, feed generation, influencer discovery, and content moderation — all built with clean class architecture and encapsulated logic.

An interactive neural network system that learns to recognize hand-drawn 5×5 pixel patterns and classify them as emojis like 😊, ❤️, or ⭐. Built from scratch using custom neuron and network classes, it features feedforward and backpropagation logic, one-hot encoding, and sigmoid activation. Users can draw emojis via console input, train the model with predefined or custom patterns, and track accuracy over time. The system includes canvas manipulation, emoji-specific training sets, engagement-based prediction confidence, and save/load functionality — all orchestrated through a modular, object-oriented architecture.

A personalized learning platform built with object-oriented design, simulating a real-world edtech system. It adapts to individual learning styles, tracks performance across lessons and quizzes, and recommends content based on engagement and topic preferences. The platform supports streak-based motivation, dynamic course and lesson management, and intelligent learning path generation — all orchestrated through a modular class hierarchy that models users, courses, lessons, and assessments.

A dynamic, full-stack portfolio website built with FastAPI and Jinja2, showcasing projects, skills, and interactive admin features. The backend leverages Uvicorn for high-performance async serving, PyMongo for seamless MongoDB integration, and Python-multipart for handling file uploads. Environment variables are managed securely using Python-dotenv, ensuring clean deployment across environments. Designed for scalability, responsiveness, and maintainability, this site reflects both technical depth and design polish.