SentinelAI

Inspiration

Critical infrastructure such as factories, hospitals, airports, power plants, and smart cities depends on continuous monitoring. Most existing systems only generate alerts after a problem occurs, leaving operators to manually analyze data and coordinate responses. We wanted to build an autonomous AI system that doesn't just detect issues but reasons about them, predicts failures, and recommends the best actions before incidents become disasters.

What it does

SentinelAI is an autonomous multi-agent AI platform that monitors critical infrastructure using CCTV cameras, IoT sensor data, and operational logs.

The platform can:

Detect fire, smoke, PPE violations, equipment damage, and unauthorized access using computer vision. Monitor sensor data such as temperature and vibration. Predict equipment failures before they happen. Assess the severity of incidents using AI reasoning. Recommend the best response and maintenance plan. Generate incident reports and notify responsible teams in real time. Display all information through an interactive monitoring dashboard.

How we built it

We designed SentinelAI using a modular multi-agent architecture where specialized AI agents collaborate to solve complex operational problems.

Technology Stack

Frontend: React, Tailwind CSS Backend: FastAPI, Python AI Models: Google Gemini API Computer Vision: OpenCV, YOLO Multi-Agent Framework: CrewAI / LangGraph Database: PostgreSQL Deployment: Docker

Each AI agent performs a specific responsibility such as monitoring, prediction, planning, reasoning, and communication.

Challenges we ran into

Designing effective collaboration between multiple AI agents. Simulating real-world industrial sensor data for testing. Balancing fast real-time analysis with AI reasoning. Integrating computer vision and language models into a single workflow. Building an enterprise-style dashboard that remains simple and easy to understand.

Accomplishments that we're proud of

Developed an autonomous multi-agent AI architecture instead of a traditional chatbot. Combined computer vision, predictive analytics, and AI reasoning into one platform. Designed a scalable solution suitable for multiple industries. Created a practical solution with strong commercial and social impact. Built a project that demonstrates how Agentic AI can proactively prevent incidents rather than simply reporting them.

What we learned

Throughout this project, we learned how Agentic AI systems coordinate multiple specialized agents to solve real-world problems. We also gained experience integrating computer vision, backend APIs, AI reasoning, and modern web technologies into a unified enterprise application. Most importantly, we learned that AI creates the greatest value when it helps people make faster and safer decisions.

What's next for SentinelAI

Our future roadmap includes:

Integration with real industrial IoT devices. Drone-based infrastructure inspection. Digital Twin support for smart factories. Edge AI deployment for low-latency monitoring. Predictive maintenance using historical machine learning models. Mobile application for field engineers. Integration with ERP, SCADA, and industrial automation systems. Support for smart cities, hospitals, airports, and railway infrastructure. AI-powered root cause analysis and autonomous workflow execution.