Projects

Selected work in disaster guidance, applied machine learning, and offline IoT monitoring.

Disaster Insight Hub Assistant

An AI-powered disaster guidance application that turns crisis reports into location-aware, context-driven responses.

Selected Technologies

PythonFlaskRandom ForestTF-IDFspaCyDocker

Project Highlights

  • Classifies crisis reports across seven categories with a Random Forest model trained on 2,345 labelled records.
  • Uses TF-IDF retrieval over a curated knowledge base to surface relevant guidance.
  • Extracts locations with spaCy and combines them with cloud LLM responses through REST APIs.
  • Publicly deployed as a Flask application.

Smart Weather Monitoring System

An offline IoT weather monitor for rural farming environments where dependable internet access is not guaranteed.

Selected Technologies

ESP8266DHT11Embedded CLCD

Project Highlights

  • Measures real-time temperature and humidity with an ESP8266 and DHT11 sensor.
  • Refreshes readings every two seconds on a local LCD interface.
  • Works without an internet dependency for practical field use.