@DavidSalas

AgroSentinel is an autonomous AI agricultural monitoring agent that detects and manages crop water stress in real time. Built for the Intel InNow Technology Fest 2026 in Costa Rica, it follows a perceive-decide-act cycle: it gathers field data (NDVI, soil moisture, temperature, rainfall, crop type, growth stage, drone imagery analyzed with Claude Vision) and enriches it with real climate data from NASA POWER. Before invoking AI, the system runs deterministic agronomic calculations based on the FAO-56 methodology (Hargreaves-Samani reference evapotranspiration, soil water balance, and real crop coefficients cited from a Costa Rican engineering thesis), so Claude reasons over validated science rather than inventing numbers. Claude then decides the exact opening percentage for a dosed irrigation valve, explains its reasoning factor by factor, and states its confidence level.