Laeldavis - Machine Learning (Level 1) Pathway

Technical Area:

Learned a lot about Git that I thought I already knew. I think for a lot of the things learned, most of them I was already somewhat introduced to, but wasn’t particularly savvy in. Colin introduced me to the different extensions about VSCode. I’ve had VSCode for awhile now and never really used it; preferred to use vim (and maybe PyCharm). However, after the 2 videos that he broke down, I completely fell in love with VSCode. The workflow seemed much more streamline and easy to use. In addition, I VSCode has a vim plugin. He gave plenty of examples as of why he preferred different methodologies and technologies over others and I don’t see that many people going into that level of detail, so that tells you about the people that teach at StemAway.

As well, I learned a lot more about NLP and BERT and how useful the transformers are and the origin of why they are even used in text.


  • Downloaded multiple tools and libraries like Spacy, Sentence_transformers and the Transformers library. For the majority of the other libraries, I already had them previously installed.

Soft Skills:

  • I learned about Gsuite and different ways to interact and communicate with teammates. There are plenty of resources and people that are very knowledgable, so I plan to utilize them and learn even more given this opportunity.
  • I learned about how to create elaborate presentations and how to follow along your viewers when you present information. All the presenters are extremely good at breaking down difficult concepts and making sure you not only understand what these technologies do, but help you understand why.


  • Followed the demo of the engineer that made a content-based recommender system and liked that project, so tried it out with a query of books that I was able to put into a dataframe and use some of the stuff we worked on like the cosine similarity and the similarity matrix. Also learned more about collapsing columns and extracting out irrelevant columns that weren’t useful being a feature to train.
  • I got the Anaconda environment working. I currently am working on an Arch Linux distribution and usually just use venv/virtualenv (forgot which one I use) for my working environments, but wasn’t aware Anaconda had a lot of what those 2 packages had and much more.
  • I got to scrape the Discourse forum a little during some of the demos that I saw. I think it’s pretty cool to interact with the HTML/Python duo for web scraping with selenium and webdriver. It’s definitely a building block that can be scaled pretty nicely into different projects.
  • I understood that PyTorch introduction pretty well after going through majority of the resources provided and revisiting the Coursera ML course that I was in a while ago. I find the mathematical portion pretty interesting because when you understand about gradients and tensors and loss/activation functions mathematically, my biggest challenge is really adjusting to the different syntax, so my challenge is to begin to read documentation longer.