Hello visitor! My name is Martin Račko and I’ve been around computers all my life. I was three years old when I’ve played the first games – Grand Prix Circuit, Prince of Persia, Tank Wars and F29 Retaliator – on a then powerful Intel 286 with a dedicated turbo button. I still remember typing “cd gprix” before I could actually read or write.

Fast forward to 2014 when I’ve decided to take my gaming one step further by trying out the other side of the coin – game development. Not having any prior coding skills, I’ve downloaded the free version of Game Maker by YOYO Games. And I loved it. The fact that I could turn lines of code into something I could interact with got me hooked. This all happened while I was reading “Viper Pilot: A Memoir of Air Combat” by Dan Hampton which had ultimately inspired me to make a SAM launcher action simulator game. Almost three years later, after countless hours of learning by doing, SAM Commander Arena was finally ready to be published.

In 2018, the concept of artificial intelligence and machine learning caught my attention. Knowing that it existed for quite some time, I wanted to dig deeper into it and understand the principles behind it – does it really work or is it just some hype? First, I had to learn how to code in Python and once I was comfortable enough with it, I jumped onto neural networks and machine learning algorithms. Taking Andrew Ng’s Deep Learning Specialization (which I fully recommend), reading a lot of literature and doing tutorials, I was quickly able to build and train my own neural networks. But most importantly, I was amazed how elegantly such AI worked. With the help of numerous interconnected “neurons” a piece of code was able to learn things on its own! It wasn’t hype, it really had a lot of potential!

Armed with enough know-how, I decided to pursue my own larger AI projects – Artificial Intelligent Value Investor (AiVi) and a Slovak to Eastern Slovak Translator. Not afraid of failure, I thought to myself, what could go wrong? Worst case, it doesn’t work and I gain better understanding of AI. After months of data preparation and manipulation, trying out different types of algorithms and lengthy hyper-parameter tuning on the AiVi, I slowly started to see some progress. The AI was able to teach herself how to identify overperforming stocks! Although knowing that the data used was somewhat limited and the final model far from finished, it was s huge personal success. A very promising AI prototype for the stock market was born. Meanwhile I was also working on the Slovak to Eastern Slovak Translator. Here, I was able to build a rudimentary dictionary, a working back-end and front-end and finally managed to deploy a functioning prototype on AWS. The major challenge with the translator currently is the lack of available data. Data has to be created more or less from scratch which is very time consuming. The development, however, continues forward.

Let’s see what the future brings. I will most certainly continue developing my current projects and I will definitely keep my eyes and mind open for any new creative ideas.

Martin Račko

Founder and CEO of Caron Studios s.r.o.