Inspiration

Organizations generate large volumes of documents, but extracting meaningful insights is often manual and time-consuming. We wanted to build an AI-powered system that can analyze enterprise PDFs and provide actionable intelligence.

What it does

Qwen Agent Society is a multi-agent RAG platform that ingests PDFs, performs semantic retrieval, and generates Research, Finance, Compliance, and Validation insights through specialized AI agents.

How we built it

We built the solution using React, FastAPI, LangGraph, Qdrant, Sentence Transformers, and PyPDF. Documents are processed into embeddings, stored in a vector database, retrieved through semantic search, and analyzed by AI agents.

Challenges we ran into

• Building an end-to-end RAG pipeline
• Optimizing retrieval quality
• Orchestrating multiple AI agents
• Managing context flow between agents ## Accomplishments that we're proud of
• Developed a working enterprise-grade multi-agent RAG system
• Implemented PDF ingestion, vector search, and AI-driven analysis
• Successfully integrated LangGraph-based agent orchestration ## What we learned
• Multi-agent AI workflows with LangGraph
• Vector databases and semantic search
• Retrieval-Augmented Generation (RAG)
• Full-stack AI application development ## What's next for qwen-agent-society
• Hybrid Search (BM25 + Vector Search)
• Cross-Encoder Re-ranking
• Multi-document analysis
• Advanced memory and agent collaboration
• Cloud deployment at enterprise scale

Built With

• built-with-python
• fastapi
• git
• langgraph
• pypdf
• qdrant
• react
• rest-apis
• retrieval-augmented-generation-(rag)
• semantic-search
• sentence-transformers
• typescript
• vite