About the Team
We are Delicephalonlar, a three-member interdisciplinary team working at the intersection of AI, risk analysis, chemical dispersion modelling, and health impact assessment. Our team brings together students from medicine, computer engineering, graphic design, and chemical engineering.
Our team name was a deliberate choice. “Cephalon” refers to the hemisphere of the brain, while “deli” is the Turkish word for “crazy.” For us, it represents bold and unconventional thinking. We chose the name Delicephalonlar because we unite medicine, chemical engineering, and computer engineering within a single integrated way of thinking. It reflects a mindset that is ambitious, methodical, and grounded in calculation.
• Lorin Türe – Risk Modelling & Impact Analysis
Medical student at Ağrı İbrahim Çeçen University Faculty of Medicine
• Oğuzhan Paksoy – AI & System Architecture
Student in Computer Engineering and Graphic Design at Tarsus University
• Gökçe Kömürcüoğlu – System Modelling & Optimisation
Chemical Engineering student and undergraduate researcher at Eskişehir Osmangazi University
Motivation to participate
We first heard about the AI4Purpose Hackathon through LinkedIn. What attracted us most was the opportunity to test our ability to define a real-world crisis problem and build a fast, meaningful solution under pressure.
The topic was important to us because emergency situations require rapid, data-driven, and reliable decision support. We see AI not as a replacement for human judgment, but as a tool that can help decision-makers respond faster and more effectively when time and uncertainty matter most.
Our AI Solution: NOXAi
During the hackathon, we developed NOXAi, a modular decision-support system for analyzing Natech risks—technological disasters triggered by natural hazards such as earthquakes.
Our solution focuses on the risk of post-earthquake chemical releases, especially in industrial areas. It is designed for decision-makers, emergency planners, and disaster response institutions that need transparent and structured risk analysis.
NOXAi combines:
• seismic damage probability
• chemical hazard coefficients
• meteorological dispersion factors
• population exposure data
to produce a risk score, confidence interval, and map-based risk zones.
AI contributes by helping optimize parameters and support scenario analysis under uncertainty. What makes the solution impactful is that it does not aim to automate decisions, but to provide calculable, transparent, and explainable risk outputs for crisis response.
Future Development
Our next step is to further validate and improve NOXAi through pilot applications in industrial regions, starting ideally with a high-risk urban area such as Istanbul. If successful, we believe the model could scale from pilot zone to city level, then national and international use.
In real-world terms, the system could support:
• industrial risk mapping
• emergency preparedness planning
• critical infrastructure protection
• integration with disaster coordination systems
To move forward, support in data access, institutional collaboration, mentorship, and pilot testing opportunities would be extremely valuable. Collaboration with public institutions, industrial stakeholders, and projects like MEDAIGENCY could help us test the model in operational settings and turn the prototype into a validated field-ready system.
We would like to warmly thank Mr. Ayman Rahmeh, MEDAIGENCY Communications Manager at the American University of Beirut, for facilitating this series of interviews.
Next Med
