-
-
Cloud Run
-
Smart Agro Fleet
-
VM Instances
-
Smart Agro Autonomous Fleet live
-
Maps Platform
-
API & Service
-
Dockerfile
-
Main.py
-
Multi-Agent Architecture
-
-
Requirement.text
Inspiration
Modern agriculture faces severe challenges regarding inefficient resource allocation and water management. We wanted to build an autonomous system that removes human delay and optimizes crop care automatically using advanced multi-agent coordination.
What it does
Smart Agro Autonomous Operations Fleet is a multi-agent automation platform.
- Data Collector Agent: Continuously gathers telemetry data such as soil moisture, ambient temperature, and crop metadata.
- Agro Analyzer Agent: Uses Gemini AI logic to evaluate risks and decide whether irrigation or soil treatment is needed.
- Action Executor Agent: Triggers physical infrastructure commands (like opening pump valves) and logs operational state.
How we built it
- Framework: Python & Flask containerized with Docker.
- Architecture: Google Agent Development Kit (ADK) multi-agent design with Agent-to-Agent (A2A) protocol concepts.
- AI Model: Google Gemini API for analytical decision-making.
- Cloud Hosting: Deployed seamlessly on Google Cloud Run for scalable serverless execution.
Challenges we ran into
Orchestrating multi-agent state management cleanly within a lightweight serverless environment like Cloud Run, ensuring robust execution flow across data gathering and action triggers.
Accomplishments that we're proud of
Successfully deploying a working autonomous multi-agent backend pipeline accessible via live HTTP endpoints from mobile environments.
What we learned
Mastered deploying multi-agent agentic workflows on Google Cloud Infrastructure using modern containerization.
What's next for Agro Autonomous Operations Fleet
Integrating real-world IoT hardware sensors (ESP32/Raspberry Pi) and enhancing Gemini models with multi-spectral satellite imagery for predictive pest management.
Log in or sign up for Devpost to join the conversation.