CV
My main focus within game development is anything more low-level related,
I love optimizing code and getting the most out of a small piece of code
or working inside of an engine and adding new things or improving the old systems.
Besides that I also love gameplay, for the freedom and creativity that comes with it.
I like working in a team where everyone can help each other and shares a passion for making games.
Currently, I'm studying at DAE. Where I'm starting my third year
and looking for an end‑of‑study internship.
I have a deep love for general problem solving and like a good challenge,
being able to make something and then afterwards being able to see it all come together and function properly,
together with being able to constantly learn new things and improve is something that keeps me going.
A picture of me, and some friends
A 8080 microprocessor emulator that can run classic space invaders and COM programs,
Fully written in C++ I mainly chose this emulator to learn more about emulator development and low-level programming.
This was a research project I did about reinforcement learning in 2D shooter games.
This is a small Unity project I made to experiment with compute shaders and ray marching.
Under construction
A brand new game I'm working on with a lovely team, currently still in prototyping phase but soon enough
more about the game and development will be available here.
Under construction
For my graduation work I'm writing my own compiler for a self made language, I'll be looking in to optimazation techniques and merging scheduling and register allocation.
The Ascend is a game project I worked on in a team, it was made in Unity and version controlled by perforce. The only requrement that we had was to make a local co-op game.
Add me on LinkedIn
Send me a mail
A complete 8080 microprocessor emulator that can run classic space invaders and some COM programs,
I mainly chose this emulator to learn more about emulator development and low-level programming.
All 255 opcodes have been implemented and tested, also some dedicated shift hardware has been emulated as space invaders use this for moving pixels on the screen and thus adds a hardware shift register to help with the math.
As the 8080 is a microprocessor and not a complete machine like a Gameboy, it does not have a standard input or sound, only some ports that can receive and send signals.
So for sending keyboard signals I used Qt as these are not really part of emulating an 8080.
For now there's no sound but emulating a sound card used in old arcade machines might be a fun project for the future.
This was a research project I did about reinforcement learning in 2D shooter games, while this project didn't turn out as well as I hoped it would, I did learn a lot from it and really enjoyed it.
The code in this project isn't the best as it was mainly focused on research. However besides these things I did get some nice results and managed to write a small article about the process.
You can find the full research paper with all the code on my github (see top left for link)
but here's my conclusion.
The aim of this paper was to see how well I can use RL to learn a bot basic AI behaviours and how well it would perform in a 2d shooter. From my attempt, I can conclude that basic AI behaviours aren't too hard to teach, but teaching an agent how to properly navigate or explore a map is a lot harder. I do believe that it is possible, if you have the required knowledge and time for it. When you do manage to do it properly, I can see various advantages for using Rl over rule-based systems such as; - Minimal code needed for the underlying algorithm in some cases - Decrease in the time spent tuning parameters for different situations - The ability to produce different bot personality types I'd say that teaching an agent how to explore a room was hard for my low experience when starting this project - combined with the limited time. It might have been better to start off simpler. However, I did learn a lot about machine learning in general and I have improved dramatically. So therefore - when I have time - I might restart or continue on this project and try to make something better. Further work could be spent on trying different environmental setups with the same agents to see how well they perform in an unknown environment. Some of the tasks have already been tested in different environmental setups like the searching for items and shooting an enemy as those setups were random. There is a clear future for RL in games, and the potential is limitless. When you combine it with more conventional rule-based systems, the options are endless. I am genuinely excited to witness the progressions.
This was my first time using unity's compute shader so I started off by making a small test project randomizing the color and position of some cubes,
I found that it wasn't too diffrent from a class I had in school where we made a raytracer and rasterizer in C++ using DirectX and HLSL,so it was suprisingly easy to get in to.
After that I felt confident enough to start on making the raymarching project.
Making the actual raymarching shader was suprisingly easy, having made a raytracer in the past, it was interesting to see the differences between
the raytracer and raymarching and seeing whats possible using the signed distance formulas raymarching uses.
If you're interested you can find and download my project and or read more about it on my github.
Using signed distance functions allows you to do some operation between shapes like masking, blending, and cutting and by combining shapes and using those operation you can create many thing's (some more usefull than others) but regardless of the use case it can be very fun to play around with.
This is a group project I'm currently working on together with some friends in school. The group consists of 3 artists and 2 programmers. In our game you play as a mushroom in a large garden, but there's a big lawnmower that keeps destroying your home, so you set out on a mission to stop the lawnmower. Currently were still prototyping so there wont be much code to show yet however I can show some things I've been working on for the game.
To lessen the workload of the artists we decided to make the animation of our insects procedural, so I made a prototype to test how well this would work and if it was a better alternative than animating the insects, the prototype was a success and I was surprised myself how easy it was to make, now there's a lot less to animate and the procedural animations allows versatile movement so our insects can seeminglessly traverse the landscape of the garden.
Next to the proceduaral animations I'm also responsible for most of the Ai, however as we are making our game in Unity there isn't too much interesting to show here, I use unity's navmesh agents combined with a Finite State Machine I wrote myself. Maybe in the future there will be something noteworthy to showcase about this.
Even though there isn't a lot to showcase here yet, this project is still on my portfolio as it's currently under construction and I am hoping to have a fun game by the end of this project.
It also shows I'm still actively working on making games next to the more scientific projects I have presented on my portfolio.
You can check out the whole progress of making the game here
The Ascend is a game where the two main characters Bling and Blong fell down a ravine and on their way out they uncover the mysteries of a long lost tomb.
It's a game where you have to work together to solve puzzles and make your way trough the level, the only catch is that you bound together by a rope, luckly you can use this rope to slove the puzzles and get out of the ravine.
Everything was made in unity and version controlled with perforce.
We were a team of 1 programmer and 3 artists. Initially, we had 2 programmes but one of them dropped out before during the prototyping phase.
So essentially all the programming was done by myself.