Computer Vision & AI

I’m interested in systems that can observe the world, interpret what they see and respond accordingly. My work in this area combines computer vision, robotics and practical AI experimentation.

Rather than focusing on AI as an abstract concept, I approach it as an engineering problem: building systems that can detect, classify and react to real-world input.

Many of these experiments originate in robotics projects, where reliable perception is essential for autonomous behaviour.

Computer Vision in Robotics

Robotics vision plays a central role in many of my experiments. Cameras provide robots with situational awareness: detecting objects, identifying markers and estimating position within an environment.

Recent work includes:

• AprilTag and ChArUco detection for robot localisation
• Vision pipelines for competition robotics
• Camera-based alignment and positioning systems
• Combining vision data with odometry for autonomous navigation

These experiments are often carried out within FIRST Tech Challenge robotics, where vision systems must operate reliably in dynamic environments.

AI-Assisted Development

AI tools are increasingly becoming part of the engineering workflow. I use AI not as a replacement for development, but as a tool to accelerate exploration, testing and prototyping.

Areas of experimentation include:

• AI-assisted coding workflows
• local AI models for development support
• generative tools for design and experimentation
• AI as a companion in complex problem solving

The focus is always on practical application rather than automation for its own sake.

Educational AI Prototypes

Another area of interest is making AI and computer vision understandable through hands-on prototypes.

Many people experience AI as a black box. By building interactive demonstrations, it becomes possible to show how machine learning and vision systems actually work.

Projects in this area focus on:

• visual machine learning demonstrations
• interactive AI experiments
• educational robotics systems
• prototypes designed for workshops and STEM education

Current Project Highlight

Train to Sustain Train to Sustain is an AI-driven waste sorting prototype that uses computer vision to identify different types of waste and mechanically sort them. The system combines: camera-based object recognition embedded control systems servo-driven sorting mechanisms real-time classification feedback The goal is to create a demonstrator that connects sustainability, AI and robotics in a tangible and understandable way. More details about this project will be published as development progresses.

Future Experiments

Computer vision and AI continue to evolve rapidly.

Current areas of exploration include:

• edge AI for robotics
• lightweight vision models for embedded systems
• educational AI labs based on Raspberry Pi hardware
• combining robotics, vision and local AI infrastructure

These experiments form part of an ongoing exploration of practical AI systems.