PARAMSINGH

Engineering Honours research project using modified ResNet50 to detect Armillaria root rot in cherry trees through multi-spectral imaging analysis (RGB + NIR bands). The system processes aerial imagery to identify early signs of disease.

This project explores the use of large language models (LLMs) for biomarker analysis and health recommendations using time series data.

Produced an iOS application that digitises handwritten construction documents using compressed large language models and adaptive image processing. The system operates completely offline and was developed for a major Australia/New Zealand construction firm to streamline on-site document processing.

Development of a two-wheeled autonomous robot utilizing custom PCB-mounted photosensors for maze navigation. The system combines analog circuit design for line detection with pathfinding algorithms implemented on a PSoC 5 microcontroller, enabling shortest-path navigation and precise speed control through projected mazes.