Machine learning is rarely a linear process that magically produces results, and iterating between your models and your data will ensure a solid foundation to build your custom ML solutions on. Lj and I created a spaCy project to showcase the functionality of the SpanRuler within a NER pipeline, but when we didn't see the improvement we were looking for in the initial pipeline evaluation, I looked into the data and found some inconsistencies in the annotations. This led me to go back with a new annotation workflow and improve the overall pipeline.

In the span of less than a year, I went from knowing almost nothing about natural language processing and machine learning to having a job at one of the coolest ML/NLP companies out there. I’m going to take a bit of time in this blog post to reflect on some of the things I did and mindsets I had that I think led me to where I am today.

I just somewhat finished the beta of a project (synsong) to practice NLP, Python, and API skills. Right now, you can input genre, popularity, and a natural language prompt (such as a sentence, quote, or paragraph), and it will create a playlist based on matches with the lyrics, song title, and artist name with the prompt.