Spaces:
Running
Running
| title: Research Assistant | |
| emoji: ๐ | |
| colorFrom: indigo | |
| colorTo: blue | |
| sdk: docker | |
| app_port: 7860 | |
| pinned: false | |
| license: mit | |
| short_description: RAG over ArXiv + open web, LangGraph powered | |
| # Research Assistant | |
| A production-grade RAG system built with LangChain 0.3 + LangGraph. | |
| - Ingests papers from ArXiv and the open web | |
| - Hybrid retrieval (dense + BM25) with cross-encoder reranking | |
| - Multi-step LangGraph workflow with query rewriting and self-reflection | |
| - Multi-provider LLM backend (HF Inference, Groq, Anthropic, OpenAI) | |
| **Source code**: [GitHub](https://github.com/AhmedMostafa167/Projects/tree/main/01-LLM-RAG/research-assistant) | |