Vector search demo with the arXiv paper dataset, RedisVL, HuggingFace, OpenAI, Cohere, FastAPI, React, and Redis.

No description available

Production-grade RAG system for arXiv papers: hybrid BM25 + vector search, local LLM inference, agentic retrieval with LangGraph, and full observability. Built with FastAPI, OpenSearch, PostgreSQL, Airflow, Ollama, Redis, and Langfuse.

Advanced Retrieval-Augmented Generation (RAG) pipeline for querying research papers from arXiv. Features include hybrid retrieval (BM25 + vector search), CrossEncoder reranking, hierarchical chunking, multi-query expansion with Reciprocal Rank Fusion, intelligent context compression, answer validation, and Redis-based prompt caching

An autonomous AI agent designed to streamline academic and professional research workflows. The system orchestrates a multi-step pipeline that intelligently routes queries between academic sources (ArXiv) and general web search (Tavily). It leverages ChromaDB for RAG-based information synthesis and Redis for robust state management.