FAISS library compiled for iOS, macOS, tvOS, watchOS

FAISS and OpenMP library compiled for iOS, macOS, tvOS, watchOS

Built a chatbot using AI& ML, LLMs , NLP, that makes conversation better on questioning about the different kind of mobiles, using Gradio for interactive interface, along with NLP ,faiss-cpu, pandas, data processing.

AI-powered Vipassana meditation guide based on S.N. Goenka's teachings. Features RAG with FAISS & BM25, MongoDB chat history, PWA support for mobile installation, voice input/output, and multilingual support (English, Hindi, Marathi). Built with Flask, OpenAI GPT, and Sentence Transformers.

A Conversational Retrieval-Augmented Generation (RAG) system that allows students to query faculty details (short names, emails, mobile numbers, and room numbers) from university PDF documents. It combines BM25 keyword search + FAISS semantic search + memory-aware retriever to provide context-aware, natural language answers.

Zabiverse (MARS Project) – A multimodal AI-powered smart campus assistant integrating a NAO H25 humanoid robot, React Native mobile app, and ReactJS web portal. Features include voice and facial recognition, LLM-based Q&A, semantic document search (FAISS), role-based access control, and file uploads for students, faculty, and management.

No description available

No description available

RAG-based customer support chatbot for TalkHome Mobile using FastAPI, Gemini, FAISS, and Streamlit.

The Jupiter FAQ Bot is a semantic-search chatbot for Jupiter’s mobile banking app, using Azure OpenAI, FAISS, and a multilingual Streamlit UI.

*VectorLite* – Sub-50 ms semantic search on-device, 90 % faster & 71 % leaner than FAISS-mobile. Custom HNSW-int8 + TFLite MiniLM-L6, 1 M docs offline, <12 MB APK. Built for Google-scale edge ML.

Fincierge is a Flutter-based mobile application featuring an AI-powered assistant built using a Retrieval-Augmented Generation (RAG) pipeline. The app combines thoughtful UI/UX design with a scalable backend powered by FastAPI, Firebase, FAISS, and Gemini AI.

The system leverages a combination of **text preprocessing**, **vector search with FAISS**, and **transformer-based QA models** to understand and answer user questions. The backend is powered by **FastAPI**, making it easy to integrate into web or mobile apps.

An enterprise-grade, unified RAG chatbot providing seamless customer support across the Reliance ecosystem (Jio Mobile, Fiber, & Mart). Powered by Gemini 3 Flash, LangChain, and FAISS for high-accuracy, multi-domain intelligence with real-time metadata tagging and a modern Streamlit interface.

AI-powered fitness coach with real-time pose detection with camera mode and multiplayer rival mode. Features 8 exercises, 3-layer form analysis (Rule+FAISS+ML), GPT-4o coaching, 2D avatar, YouTube music, mobile remote control and modern React UI with dark theme.

Zoikon is a real-time AI-powered chatbot developed for Zoiko Mobile UK. It provides intelligent menu-based navigation, voice interaction, website redirection, and feedback collection with database storage. This project demonstrates full-stack development using FastAPI, MySQL, RAG (FAISS), and a modern frontend UI.

A personal one-page portfolio website with a built-in chatbot powered by Retrieval-Augmented Generation (RAG). The chatbot is trained on my resume and built using LangChain, FAISS, and Groq’s LLM. The project is fully deployable on Render with a FastAPI backend and a mobile-friendly HTML frontend

A mobile app that lets users find visually similar fashion items just by taking or uploading a photo. It uses deep learning (CNN/ANN) to extract image features, FAISS for fast similarity matching, and combines a React Native frontend with a FastAPI backend to deliver accurate, real-time fashion recommendations.