I learned that there is toooooo much to learn. Either tackle the software, math, or the ML. Currently tackling the Software, short term goal is to get very good with programming. Then I will dive into the mathematics and domain knowledge of finance.
Biggest challenge during research was that I don’t have enough financial domain knowledge to help me verify what I was doing well in the ML steps and not enough ML intuition to tune. After a while it sort of fell apart. But that’s good because it’s clearer now what I want to learn and the steps to learn it.
I also used to have a repo that allowed a collaborative environment and shared ML models to subclass from and create strategies but I’ve since deleted the repo and can’t find it on my laptop (think it’s on my Desktop somewhere). Will update the link once I find it.