-
-
Sistam Architecture
Inspiration
Traditional agriculture faces extreme climate risks—unpredictable heatwaves and rapid soil moisture depletion—while photovoltaic solar farms lose significant power efficiency when panels overheat. AgroVoltaic-Edge was born to solve this dual challenge by creating an autonomous Physical AI Edge Framework that synchronizes smart micro-irrigation with dynamic solar canopy microgrid cooling.
What it does
AgroVoltaic-Edge is an end-to-end autonomous agent system that continuously monitors field microclimate telemetry (soil moisture, ambient temperature, sunlight intensity) and leverages large language model reasoning via Nebius Token Factory and NVIDIA GPU infrastructure. When heat thresholds are breached or soil dries up, the system automatically triggers micro-drip irrigation and optimizes solar panel tilt angles in real time.
How we built it
- Backend & API: Python, Flask, and Gunicorn deployed on Render Web Service.
- AI & Reasoning: Nebius Token Factory API utilizing high-performance open models (
Qwen3-Coder-480B-A35B-Instruct,Kimi-K2-Instruct,gpt-oss-120b) with an edge rule-based fallback logic. - Uptime & Telemetry: UptimeRobot 24/7 keep-alive monitoring and real-time CSV telemetry logging.
Challenges we ran into
Integrating real-time edge telemetry with low-latency LLM decision loops required building an automatic multi-model fallback mechanism to ensure uninterrupted operation even during transient network or cloud API latencies.
Accomplishments that we're proud of
Successfully deploying a 100% active zero-downtime Physical AI agent framework on Render cloud that provides real-time REST API telemetry for both smallholder farmers and industrial agro-hubs.
What's next for AgroVoltaic-Edge
Scaling the physical sensor hardware deployment across BSCIC smart agricultural facilities in Bangladesh and aligning with national Vision 2030 smart industrial targets.
Log in or sign up for Devpost to join the conversation.