Velen Sight: Machine Learning for Hearthstone
Technologies used: Python, ML, AWS, Docker, Qt
Velen Sight is an ML-powered game assistant for the popular online competitive card game Hearthstone. It sits passively alongside the game and watches the game for cards being played, then feeds this information into its learning models to determine the likelihood that your opponent will also play other cards.
I designed and developed the app and sold it on Gumroad as a recurring subscription service. The subscription paid for periodic updates to the deck database, so that the predictions stayed accurate with the decks that the community was currently playing.
Streamlined design
I put as much effort into UI/UX as I do backend engineering, and Velen Sight is no exception. Making UIs clean and unobtrusive—while remaining intuitive—is something I take a lot of pride in. I believe that a good UI/UX makes a difference in making your product stand out.
Full stack development
GUI design, ML model building, data scraping, cross-platform development, microservice architecture, deployments, live updates— I did it all for this one. All scripted, automated, and clearly documented, every facet of the system was written to be efficient and robust so that I could manage the stack easily and iterate quickly.
Deployments were simple. Updates of the models were seamless. The system was designed from bottom to top to be lean and easy to manage.
Take a look at the code here (TODO)
