AI-powered Indian legal assistant using RAG — 164K chunks, hybrid search, cross-encoder reranking, Gemini 2.5 Flash streaming, 3D Next.js UI.

Production-ready semantic vector search for Django — searches across FK, M2M, and reverse relations by traversing your model graph. Pluggable backends: ChromaDB, FAISS, Qdrant.

Serverless Retrieval-Augmented Generation system on AWS using Bedrock Titan embeddings, Aurora PostgreSQL pgvector, Lambda, API Gateway, SageMaker, and Streamlit.

RAG Mini Project — Retrieval‑Augmented Generation chatbot with FastAPI backend (Docker on Hugging Face Spaces) and Streamlit frontend (Render), featuring document ingestion, vector search, and LLM‑powered answers

RAG-based chatbot that ingests Healthline URLs to produce concise, source‑grounded summaries & answers. Built to eliminate manual copy‑pasting & streamline insight extraction: paste article links, ask a question, & get reliable, Healthline‑only results fast.

Advanced RAG with hybrid search, query classification, answer fusion, and self-correction

AI Document Intelligence System for deep analysis and semantic querying of ingested docs and other pdfs

A WhatsApp bot with Baileys + Python RAG catalog search + Ollama/Mistral for cha

My First RAG pipeline using local ollama models

CampusAI is an agentic RAG assistant for university queries—routes questions to academics/notices/PYQs, retrieves context from MongoDB, and answers via LLMs (FastAPI backend).

Gemini-powered, zero-hallucination AI chatbot for Montfort ICSE — semantic search, IUI Engine v2.5, verified school data, and hybrid RAG architecture.

RAG PDF chatbot, retrieval-augmented QA over PDFs using FAISS and Ollama Llama 3.2:3b.

🕵️‍♂️ Vibe Checker — AI-платформа для поиска мест по атмосфере («вайбу»). Анализ отзывов через LLM (Gemini), семантический поиск (Qdrant) и детальное сравнение заведений. Стек: FastAPI, Next.js, PostgreSQL, Redis.

Lean vector embedding provider package for CitOmni with unified contracts across providers and profile-based adapters.

Swarm-A2P implements SwAPTP, a protocol for autonomous agents to resolve natural-language content intents against peer-to-peer file swarms, without a central orchestrator, a semantic layer.