The recent wildfires in Los Angeles County in January 2025, which affected thousands of lives and caused substantial property damage, have underlined the urgent need for new, advanced, data-informed strategies for efficient and proactive management of such devastating events. However, the development of such data-informed solutions is still largely hampered by limited access to multisource, multi-resolution remote sensing imagery, leaving many in the machine learning and computational sciences communities unable to contribute robustly. This project uses large language models (LLMs) to extract properties and relationships from relevant LA fire data sources, storing them in a comprehensive knowledge database. By integrating complementary wildfire-related information, the framework facilitates monitoring of key physical parameters, such as real-time evacuation orders, meteorological variables, and air quality indicators. The project aims to develop an LA Fire Knowledge Graph-Agent (LAFireKG-Agent) platform--an autonomous and end-to-end LLM-based framework designed to meet the diverse data needs of end users, and enhance situational awareness for both safety and timeliness in wildfire risk management. The LAFireKG-Agent framework focuses on three key objectives: rapid decision-making, predictive modeling, and complex reasoning. Beyond these core capabilities, end-users, including computational scientists, environmental scientists, and risk managers, will be able to explore