Building Needflare: An Autonomous Disaster Intelligence & Logistics Agent with Gemini 3.7, Gemma 4 & Google Veo
This article was created for the purposes of entering the Google Cloud: All Things Agentic Hackathon. In catastrophic natural disasters like flash floods, hurricanes, and earthquakes, cellular networks collapse and emergency dispatchers are overwhelmed with thousands of fragmented, panic-driven field messages. Volunteers trying to coordinate over open radio frequencies or mesh airwaves risk…
Needflare is an autonomous disaster intelligence and logistics agent built using Google Cloud and cutting-edge AI models such as Gemini 3.7 Flash, Google Gemma 4, and Google Veo 3.1. The system addresses the challenges of communication collapse and information leakage during catastrophic natural disasters like flash floods, hurricanes, and earthquakes.
Needflare's architecture consists of three decoupled tiers: Edge Tier for volunteer terminals, Asynchronous Ingestion Tier using Google Cloud Pub/Sub, and Autonomous Taskmaster Tier powered by Google Cloud Run and Gemini 3.7 Flash. The system ensures privacy by scrubbing PII on-device using Google Gemma 4 and falls back to on-device heuristics if cellular or satellite signals are lost.
Upon network recovery, buffered reports are burst-synchronized, and an asynchronous taskmaster evaluates crisis severity, computes humanitarian aid payloads, and auto-dispatches logistics convoys without human intervention.
Brief written by urgent.news from Dev.to's own syndicated text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
