All Projects

Decentralized photo‑sharing app with layered architecture and modern DevOps practices.

• • Clean Handler–Service–Repository backend in Go, ensuring a modular, and maintainable codebase.
• • Containerized the application and PostgreSQL database with Docker Compose for seamless deployment on DigitalOcean.
• • Built secure auth with JWT/bcrypt, added SendGrid email verification, and integrated IPFS for decentralized photo storage.

Content‑addressable storage with chunk‑level deduplication and concurrent pipelines.

• • Goroutines and channels implemented for concurrent upload and retrieval pipelines, improving throughput and responsiveness.
• • PostgreSQL used as a metadata store to manage file chunks, indexes, and deduplication logic for efficient storage.
• • Designed a modular Handler–Service–Repository architecture to improve maintainability, and enforces separation of concerns

Research‑backed API optimized for asynchronous performance over large anime datasets.

• • Endpoints for search, genre, studio, and user AnimeLists.
• • Performance: connection pooling, Redis caching, Promise.all batching.
• • Findings published at ICCSCI 2025 (Elsevier).

API for personalized K-Drama recommendations using content similarity.

• • Content-based recommendations using synopsis, genre, cast, and tags with confidence scoring.
• • Results highlights shared attributes like genres and tags.
• • Fast REST API with Node.js and Express backend.

Full-stack event management platform built as a guided course project.

• • Frontend with Next.js + Tailwind.
• • Backend APIs with Express + MongoDB.
• • Full-fledged functional web app, with admin dashboard
• • Simulated payment flows and user authentication.

Full-stack Web App to add friends, and have real time conversation

• • Frontend with React (JSX), Tailwind, DaisyUI, Zustand
• • Backend APIs with Express + Node.js and MongoDB.
• • User authentication with JWT and real time communication with Socket.IO

Web-based Flask powered Python ML to help new WAVs identify tripleS members and sub-units!

• • 3rd-Semester Artificial Intelligence Group Project
• • Frontend: HTML5, CSS3, JavaScript (with AJAX)
• • Backend: with Python + Flask
• • Machine Learning: MTCCN, cv2, torch, numpy, inceptionresnetv1, cosine_similarity

AI-Powered Health Insights: Preliminary Alzheimer's Detection Web App

• • 4th-Semester Data Analytics Group Project
• • Frontend: React (JSX) with Tailwind + React Bits
• • Backend: Flask (Python), RESTful API design
• • Machine Learning: Scikit-learn (for model training & .pkl export), Pandas (for data preprocessing)

Phi4-Mini based Locally Run Customized LLM, powered by Ollama and ElevenLabs for Streaming TTS

• • Fully customized AI Sidekick with a relaxed helpful tone inpired by Jarvis and Miss Minutes
• • Can play music on Spotify with a voice command
• • Uses Ollama (Phi4-Mini) and ElevenLabs for Text-To-Speech

Web Scraping tool to find and download images (for personal non-commercial use)

• • Scrape images from Pinterest based on a specified search query
• • Selenium for web automation and BeautifulSoup for parsing HTML. The script scrolls through the Pinterest search results and downloads images to a designated folder.
• • For personal/educational usage only, and commercial use is prohibited :D