Technicolor Waterfall by an AI Artwork

As a data scientist, I've been working with AI and machine learning technologies for quite a while. While there's no shortage of interesting problems to encounter as a data scientist, there's very little work, at least in my experience, that's creative in the same sense as creating art or music is. That's always been a little disappointing to me, so I decided to fix the problem. My solution was to start building AI systems that generate artwork. To make these systems work, I make an initial decision about the general style I want the system to learn as well as the subject matter to focus on. The computer then effectively learns to combine a style with a subject and it generates thousands of images per combination of style and subject. It's up to me to sort through those results, pick out any images that I'm particularly drawn to, and put a few finishing touches on them. In a sense, the AI system is just as much the artist as I am. I function in a role somewhere in between that of a guide and that of a curator. Being in that sort of hybrid role provides me with an interesting means of blending a technology I find fascinating with subject matter I'm drawn to. I use the resulting capabilities to explore topics related to nature, science, and technology.