MINDSAFE AI

MINDSAFE AI is an intelligent risk management platform designed for African industrial sectors. The solution combines IoT sensors, Artificial Intelligence (AI), Open Data, and a conversational assistant to monitor industrial sites, detect hazards, and predict potential incidents in real time.

The project was inspired by the lack of digital safety tools in African extractive and manufacturing industries. Many companies face serious health, safety, and environmental (HSE) issues but lack real-time monitoring and predictive analytics to prevent accidents. MINDSAFE AI was created to close this gap and make industries safer, smarter, and more sustainable.

Developed a React-based dashboard to visualize risk zones, incidents, and worker health data.

Integrated IoT sensors to collect temperature, gas levels, and worker vitals.

Implemented AI-powered risk analysis for early detection of anomalies.

Used Open Data (industrial incident records, weather data, geographic risks) to enhance predictions.

Built a Large Language Model (LLM)-driven assistant to answer safety-related questions and provide natural language recommendations.

What We Learned

How to integrate multiple data sources (IoT, Open Data, AI models) into one cohesive platform.

Handling real-time data visualization and AI-driven predictions.

Designing a system that can work offline in remote industrial areas and synchronize when network coverage is available.

Building a user-friendly safety assistant for non-technical field operators.

Challenges We Faced

Data accessibility: Industrial safety data is often private and not publicly available.

Connectivity issues: Ensuring the system functions in regions with limited internet coverage.

Scalability: Designing a solution that can handle large-scale industrial operations while remaining cost-effective.

LLM integration: Adapting AI models to generate context-aware safety recommendations in real time.

Built With

• docker
• firebase-firestore-ai-&-machine-learning:-openai-api-(llm)
• firebase-hosting-authentication:-firebase-auth-dev-tools:-github
• frontend:-react.js
• geolocation-datasets-mapping-&-visualization:-leaflet.js
• node.js-database:-postgresql
• openstreetmap-cloud-services:-aws-(ec2
• s3)
• tailwind-css-backend:-python-(fastapi)
• tensorflow-(risk-prediction-models)-iot-integration:-mqtt-protocol-for-sensor-data-streaming-open-data-apis:-industrial-incidents-datasets
• weather-api