Improve RAG Pipeline Accuracy Using OpenAI and LangChain

We are looking for an AI Engineer to improve the accuracy of an existing RAG pipeline. The system uses OpenAI, LangChain, embeddings, and a vector database to retrieve answers from documents. The issue is that some answers are not accurate or the wrong document sections are being retrieved. The task includes improving chunking, retrieval settings, prompt handling, and testing answer quality. Required experience: • Retrieval Augmented Generation • OpenAI API • LangChain • Embeddings • Vector databases • Pinecone, FAISS, ChromaDB, or pgvector • Semantic search • RAG evaluation