From Satellite to Signal: Building a Geospatial Agentic AI Stack on AWS

A hands-on workshop with AWS, Felt, and Wherobots Build an end-to-end agentic geospatial pipeline in a single day. Bring your laptop, leave with a working pattern for turning raw satellite imagery and weather data into interactive risk dashboards, driven by natural language at every stage. This in-person workshop walks you through building an end-to-end geospatial AI pipeline. Raw satellite imagery and weather data go in, move through agentic data engineering and risk scoring, and come out as interactive map dashboards. You use natural language at every stage. What You'll Learn By the end of the workshop you will be able to: Use the Wherobots MCP server as an agentic interface for exploring and processing geospatial data catalogs Build a medallion data pipeline (Bronze, Silver, Gold) using spatial SQL, raster zonal statistics, and KNN spatial joins on Apache Sedona Score physical assets for natural hazard risk using wildfire, flood, and severe weather data [see note below on verticals] Orchestrate Felt MCP tools from a Strands agent powered by Amazon Bedrock, building interactive maps from natural language prompts Build multi-altitude visualizations with Felt Style Language (FSL) zoom ramps, including H3 hexbin heatmaps that crossfade into individual buildings Connect Aurora PostgreSQL to Felt as a live data source, so maps stay current as the underlying data changes Who Should Attend Data engineers, ML engineers, and solutions architects working with physical-world data. Geospatial teams evaluating how agentic workflows layer onto existing GIS infrastructure. After the Workshop Stick around for a happy hour after the workshop. More details soon. About Wherobots Wherobots is the AI Context Engine for the physical world, built by the original creators of Apache Sedona. An AWS Partner, the platform enables planetary-scale spatial compute and Earth intelligence, letting developers build and scale AI workflows on physical-world data using the SQL and Python they already know. Learn more at wherobots.com.