Welcome to kristakorinke.com! Contact me at kristakorinke@gmail.com.
DisasterConnect is a prototype AI system designed to help people quickly access reliable emergency information during natural disasters and crisis situations. The project explores how artificial intelligence can be used to retrieve critical information from trusted sources and deliver it in a clear, actionable format when time and clarity matter most.
During emergencies, individuals often struggle to locate accurate and up-to-date information across multiple websites and government resources. DisasterConnect demonstrates how an AI-powered retrieval system can streamline access to emergency guidance, safety procedures, and available resources.
The goal of the project is to explore how AI-driven knowledge retrieval systems can improve public access to critical information during high-stress situations.
During disasters such as wildfires, earthquakes, hurricanes, or severe storms, individuals often face several challenges:
Information is scattered across multiple government and emergency websites
People may not know which sources are trustworthy
Critical instructions are often buried in long documents
Stress and urgency make it difficult to interpret complex information quickly
These barriers can slow down decision-making and reduce the effectiveness of emergency response guidance.
DisasterConnect uses artificial intelligence to retrieve relevant information from trusted emergency resources and present it in a concise, easy-to-understand format.
The system enables users to ask natural language questions such as:
“What should I do if there is a wildfire evacuation near me?”
“How do I prepare an emergency kit?”
“Where can I find local disaster shelters?”
The AI system analyzes the question, retrieves relevant information from the knowledge base, and generates a response that highlights key guidance and resources.
This approach demonstrates how AI can help transform large collections of emergency information into accessible, actionable guidance for individuals during crisis situations.
• Natural language question answering for emergency preparedness and response
• Retrieval of relevant information from trusted disaster preparedness sources
• Summarization of complex emergency guidance into clear instructions
• Context-aware responses tailored to the user’s situation
• AI-assisted access to emergency resources and safety recommendations
DisasterConnect uses a modular AI architecture designed to combine knowledge retrieval with language model reasoning.
The system includes the following components:
User Query Layer
Users submit questions related to disaster preparedness or response.
Knowledge Retrieval Layer
Relevant documents and emergency resources are retrieved from the system’s knowledge base.
LLM Reasoning Layer
A large language model analyzes the retrieved information and generates a clear response grounded in trusted sources.
Response Generation Layer
The system presents summarized guidance and actionable information to the user.
Python
LangChain
Vector database for knowledge retrieval
Large Language Models (LLM) for reasoning and response generation
Flask for prototype application interface
SendGrid for automated communication capabilities
• Retrieval-Augmented Generation (RAG) system design
• AI-driven information retrieval and summarization
• Practical application of generative AI for real-world problems
• Knowledge-based AI system architecture
• AI-assisted decision support in emergency scenarios
Potential future improvements to the system include:
Integration with real-time disaster alert systems
Location-aware emergency guidance
Multi-agent AI workflows for resource coordination
Multilingual support for broader accessibility
Mobile-optimized interface for use during emergencies
The long-term vision for DisasterConnect is to explore how artificial intelligence can improve access to critical information during emergencies and support more informed decision-making when individuals need guidance the most.
The project serves as a demonstration of how AI systems can be designed to support public safety and disaster preparedness through intelligent knowledge retrieval and clear communication.