In response to the growing complexity of cyber security threats, threat hunting has become an essential proactive security measure. However, its adoption in security operations programs is often limited to larger organizations due to the myriad of resources required to support the activity. Transformer-based Large Language Models (LLMs) present a new opportunity to democratize, automate, and enhance cyber security operations. This thesis seeks to contribute to this space in three ways: First, develop a demonstration of an LLM's ability to automate aspects of threat hunting. Second, produce a dataset that will assist with fine-tuning or training. Third, contributing to the development of a Retrieval Augmented Generation (RAG) system within AIThreatTrack.