Julia, Machine Learning, and Finance

Hello. I’ve previously asked about Julia and Finance. I’m currently going through one of the readily available introductory books on Julia. I came across this click-baity video from Siraj Learn Machine Learning in 3 Months (PyTorch Curriculum) - YouTube

So before watching this I was going to try to make a LSTM with multiple inputs correlated to what I’d like to predict, but now it seems that LSTM is outdated and people are using Transformers. My goal is ultimately to take a bunch of securities or indexes that I think are correlated with what I’d like to trade and make a predictive model or algo. So I was wondering if I can follow his curriculum in Julia and what the equivalents are or if I should just do everything in python? Thanks.

If you’re not in a rush to get something into production, I’d recommend implementing from scratch as much as possible rather than rely on pre-built machine learning libraries. You’ll learn much more and develop a better intuition building things up from a simple model. Julia is an excellent language for that.

After you’re comfortable with the basics, you may want to use something fancy that’s already built in python, but generally successful quantitative trading is done with fairly simple statistical and machine learning models, not the latest and greatest deep learning.