My interest in AI came around 2010 when I ran into a person that had an AI startup that was angel funded by a university. It was very very mathy and complex and although it was super interesting, that wasn’t going to be something I was going to pick up anytime soon, plus all that math that I didn’t know.
Around 2017 I started getting into computer vision and stuff available on AWS- Sagemaker and etc. I worked on that for several years until the barrier to entry became much lower when Nvidia GPU’s became powerful enough to start doing my own deep learning at home. The GPU cost then and now on cloud providers are quite high as a hobby or even a semi-professional interest, at least for me. When I got a few high powered GPU’s at home then I really started working a lot more with training and LLMs, transformers, and etc. I mostly don’t develop but have spent a lot of time building proof of concepts just for my own interest and some for work. Most of the stuff I use is off the shelf Hugging Face or Github built repos that I tune or otherwise hack into what I want it to do. It can be slow without expert programming skills but I do have a fairly well trained model that helps me with my crappy coding or I work with employees to get through it.
I like to stay technical but have many years of management experience. I really enjoy working with people and personalities. I have my own LLC where we do lots of work in the infrastructure/design/auditing and most recently, working in AI/ML with startups (there are a lot!)