My interest in ML/AI/Data Science began shortly after I completed my graduate work. I read an article in Scientific American on neural networks, and the idea of applying them to forecast stock prices immediately popped into my head. After reading a few books I felt like I was ready to build something. What followed was a 10-year project that produced a six-year run of annualized returns over 25%. Without realizing it, I was doing data science before data science was really a thing, and without the benefit of the Python/R code that is readily available today. I was writing code to download and process data, and then dump it in a database. I "invented" one-hot encoding not realizing it had a name. I was feature engineering and grid searching. I incorporated a genetic algorithm I thought up. I had more ideas than time to implement them. Perhaps not coincidentally, my market-beating returns dwindled to market-matching returns about the same time ML and data science were becoming mainstream. I abandoned neural networks and began searching for the next "big thing". That led me to explore discrete wavelets and statistics for a time. About ten years ago I migrated to Python, recognizing it as the best platform for exploring the many ML algorithms (XGBoost, SVM, KNN, Deep Learning, etc.), and it was around that time that I realized that what I was doing and interested in fit under the heading of "data science". As a life-long learner, I am passionate about learning new concepts and algorithms, and exploring their capabilities. I have worked on projects involving NLP/sentiment analysis, deep neural networks and transfer learning, forecasting Alzheimer's using KNN, and image classification using discrete wavelets. Some of these projects are available at github.com/jcmiii. But each time I'm exposed to a new idea, in the back of mind I'm thinking "how could this be applied to the stock market". ;)