loading
$GET/api/site-setting200
$GET/api/page200
$GET/api/skill-categories200
$CDNresolving media assetsok
$rendering page blocks
Explore the projects I’ve built across web applications, backend services, Generative AI, APIs, databases, and developer tools. Each project reflects hands-on experience with different technologies, practical problem-solving, and the process of turning an idea into a working application.
A personal portfolio platform built with a headless CMS and an AI assistant that answers questions about my skills, projects, experience, education, and technical background. The website is fully CMS-driven using Strapi, with pages and sections managed from the admin panel instead of being hardcoded in the frontend. Built the platform with Next.js, Strapi, FastAPI, and PostgreSQL, with the profile content also indexed into a vector database for the AI assistant. Implemented a RAG-based question-answering system using LangChain, LangGraph, embeddings, and vector search. The assistant combines structured profile data with relevant vector-search results to generate context-aware answers and streams responses back to the user.
A privacy-focused AI document question-answering application that allows users to upload PDF documents and interact with their content through a conversational interface. Built a RAG workflow using LangChain and LangGraph to process documents, generate embeddings, store them in a local ChromaDB vector database, retrieve relevant document content, and generate context-aware answers using language models. The application also maintains conversation history to support follow-up questions. This project was my first hands-on implementation with LangChain and LangGraph, working with Hugging Face, Google Gemini, Mistral, and OpenAI models.
A machine learning-based chatbot that determines an individual’s Prakriti type through a conversational assessment. The chatbot asks a series of questions about the user’s body and mind, analyzes their responses to determine their Prakriti, and provides diet recommendations based on the result. The project uses a Feedforward Neural Network (FNN) for Prakriti classification and combines machine learning with a conversational interface to make the assessment more interactive and accessible.
A full-stack blogging platform built using React for the frontend, FastAPI for the backend, and MongoDB for data storage. Built features such as Markdown-based blogging, user authentication, JWT authorization, email verification, blog search, tagging, user profiles, thumbnail uploads, and Cloudinary integration. The project gave me practical experience in building a complete blogging application with separate frontend and backend services.
Explore my Experience to see how I apply these skills in professional work, or follow my Timeline to see how my development journey has evolved.