This document presents a Bengali Information Retrieval System (BIRS) leveraging natural language processing techniques, especially focusing on lemmatization through algorithms like trie and dictionary-based search. It integrates various components like a corpus of Bengali root words, stop words, and anaphora resolution using Hobbs' algorithm, achieving an impressive accuracy of 97.22% in answering 19,335 generated questions. The system aims to efficiently retrieve relevant information from a large dataset, addressing challenges such as word mismatch and synonym usage.