AI and GIS for Flood Forecasting - From Data to Decisions
Speaker Bio
Tomasz Fidecki is CTO and co-founder of u11d, an expert in systems architecture and DevOps process automation. For years, he has specialized in building scalable cloud-native solutions and implementing modern Infrastructure as Code (IaC) practices using Terraform and AWS, among others.
As a practitioner and technical leader, he focuses on creating stable Continuous Integration pipelines, which form the foundation for large e-commerce systems and cloud applications. Tomasz is known for his analytical approach to solving complex infrastructure problems and sharing knowledge on technical blogs (including u11d tech blog and DEV Community). On a daily basis, he combines strategic technology management with a passion for optimizing IT system performance and reliability.
Talk Abstract
Recent advances in spatial data availability and the development of artificial intelligence methods open new possibilities for supporting hydrological forecasts and decision-making related to flood risk. This is particularly relevant in areas where classical modeling approaches are difficult to apply or require supplementation with additional data sources.
In this presentation, we will show how AI methods, combined with meteorological and geoinformation data (including digital elevation models and hydrographic networks), can support flow forecasting, water level predictions, and assessment of potential flood extents. GIS environments play a crucial role here, serving as a space for data integration, results analysis, and interpretation by experts.
Using examples from pilot analyses, we will present an approach aimed at supporting decision-making processes, anomaly identification, and better understanding of forecast uncertainties. The talk will highlight the importance of data quality, interdisciplinary collaboration, and the potential for further development of AI solutions as a supplement to existing hydrological tools.
When: April 24, 2026, 16:30 - 17:15 Language: Polish