Published optimisation research + real systems

Problem

A TSP-like logistics optimisation problem (POP) under uncertainty — the kind where standard methods stall.

Approach

I designed an AI-driven optimisation approach, adapting reinforcement-learning ideas. Alongside the research I built a RESTful system for on-device image recognition (Android app + web service + a compute server running CNNs) and a 3D drone-trajectory visualisation tool (ROS + Qt).

Result

Published a paper — and, just as importantly, demonstrated end-to-end skill: from a research idea to real, running systems.

Screenshots coming soon.