Q&A inteligente sobre o Livro Amarelo — explore propostas e ideias para o futuro do Brasil.

Multi-agent AI research workspace featuring document RAG, real-time web search, and create/edit files directly through chat. Built with Django, LangGraph, Pinecone, and AWS S3.

A travel assistant made to answer travel related queries. It responds using data retrieved from both Pinecone and Neo4j. Currently only works for Vietnam

EduPilot is an intelligent course assistant built for Indiana University students. It answers questions across four graduate courses — AML, ADT, STAT, and LLM — using a seven-stage multi-agent RAG pipeline that retrieves directly from course lecture slides and materials, cites every claim and verifies its own answers, and refuses to hallucinate.

Chatbot developed during Tata Steel Summer Internship, uses live IoT Telemetry data for various machines and allows users to query and fetch details for the same.

RAG chatbot built with n8n, Mistral AI & Pinecone — semantic search on COVID dataset with no-code pipeline

This is a production deployed RAG system that answers medical questions grounded in verified clinical documents — not hallucinations.

smolchat is a full-stack agentic AI platform (LangChain + Next.js) with multi-LLM chat, long-term cross chat memory, STT & TTS, RAG (multi-format doc support) backed by Pinecone VectorDB, and client-side chat persistence via IndexedDB. The smolchat agent integrates with tools like YouTube search, Web Search, Web Page Search, and Wolfram Alpha KE.

An AI-powered web application that lets users upload PDF documents and have intelligent conversations with them. Ask questions and get instant accurate answers with page references.

AI-powered meeting intelligence platform with transcription, RAG-based chat, action item extraction, Slack integration, and semantic search using OpenAI, Pinecone, FastAPI, and Next.js.

This is a complete, scalable Retrieval-Augmented Generation (RAG) backend application. It allows users to register, create isolated workspaces (tenants), upload PDF documents, and query those documents using natural language.

A Retrieval-Augmented Generation (RAG) pipeline built to query and analyze state-level energy policies.

Agentic RAG chatbot with hybrid retrieval, multi-tool ReAct agent, cross-session memory, and LLM-as-judge evaluation. Built on Pinecone and Claude.

Smart Photo Organiser is a full-stack face recognition web app that identifies people in group photos. Register faces with a name and photo — then upload any group image and the app detects every face, matches each one against the database using vector similarity search, and returns the top 3 closest matches with confidence scores and a image

PDF Intelligence : A production-quality React frontend for a Retrieval-Augmented Generation (RAG) API. Upload PDFs, ask questions in plain text, and get AI-generated answers grounded in your documents — complete with source citations.

A AI-powered personal portfolio chatbot built with the MERN stack and RAG architecture, using OpenAI embeddings and Pinecone to answer context-aware questions about my skills, projects, and experience.

Built a multi-agent AI Assistant using LLMs, LangChain, and LangGraph with RAG-based document retrieval. Enabled context-aware Q&A over PDFs using Pinecone vector DB. Implemented FastAPI backend with JWT authentication for scalable, secure AI interactions.

RAG pipeline to query and analyze software engineering case studies

Next.js v16, AI Applications with Gemni (Gemini 2.5) and OpenAI (GPT-5.0)

An enterprise-grade code intelligence platform that automates PR reviews, visualizes architectural dependencies, and actively generates self-healing patches using advanced AI engines and context-aware RAG.