Software & Game Developer
Code worth shipping.
I build web platforms, machine learning tools, and games, in C#, .NET, and whatever else the problem calls for.
# Dreamer, researcher, developer.
What I do
Software Developer and Team Lead, mostly in the .NET ecosystem, ASP.NET MVC, C#, and whatever JavaScript a project needs. Computer Engineering graduate, European University of Lefke.
Where I experiment
Small machine learning tools in C# and Python, sentiment analysis with ML.NET, an open-source Naive Bayes library, and from-scratch implementations like K-Nearest Neighbor.
What I build for fun
Two games built solo in Unity, both published and live on Steam. Game dev pulls on a different shape of problem-solving than web and tooling work usually demands.
Outside the editor
Photography, some of it on 500px. Keeping a balance between building, researching, and everything else is most of the point.
Featured work
A handful of the projects I have shipped, web platforms, internal tools, an Android app, and a Unity prototype.
Dynasty Protocol
A real-time strategy game built in Unity, command space fleets, mine asteroids, and expand a colony across a star system.
- Unity
- C#
Nexus GM
Nexus GM is the offline-first command center for independent creators. Never lose a design decision, forget a half-finished mechanic, or lose track of your studio's health again.
Ellipses Create Labs
A one-person studio shipping real things, a space strategy game on Steam and a dev tool for indie devs who are done with spreadsheets.
Chaos
A game prototype built in Unity, exploring procedural and physics-driven gameplay mechanics.
- C#
- Unity
Want to see everything I have shipped?
A dozen projects spanning CMS platforms, internal tools, an Android app, and a Unity prototype.
Recent articles
Mostly machine learning and .NET, Naive Bayes, K-Nearest Neighbor, sentiment analysis with ML.NET, and the math underneath all of it.
LLMs aren't always the answer. Learn why engineers over-engineer with AI when classical ML models are faster, cheaper, and more accurate for most real-world problems.
Part 3 of the series: a Windows Forms app that fetches real reviews and color-codes them using the trained ML.NET sentiment model.
Part 2 of the series: a console app that prepares the IMDB dataset and trains the sentiment model built in Part 1.
Part 1 of a three-part series: training an ML.NET model to classify IMDB movie reviews as positive or negative.
How the Naive Bayes algorithm works, why it performs surprisingly well despite a naive assumption, and a from-scratch C# implementation for text classification.