And we're back in action!
Hello, fellow datanistas!
It’s been a half-year hiatus from the newsletter, which was a necessary and welcome break for me. But we’re back! And resuming regular, ahem, programming.
In this edition, we’re going back to basics. Here are some articles and tools from my random walk on the interwebs that I’ve found beneficial to myself, and I hope they’ll do the same for you too :).
VSCode on iPad with Blink
This is one of the best implementations of VSCode on the iPad. I accidentally found this by trying to open up a GitHub Codespace in Safari — instead, Blink opened up! For the most part, the experience is seamless. I get the advantages of iOS, which is a singular focus on a single app while still getting a full VSCode experience, complete with a working Jupyter notebook extension!
Open Source Go links!
We recently got go-links at work, which are super handy! In a world where everything is addressable with a URL, getting a deep link directly to a resource is an incredible time saver and productivity hack!
Formal Algorithms for Transformers
This is another goldmine educational resource on Transformers, released by Mary Phuong and Marcus Hutter of DeepMind. This should be another helpful resource for those who are more theory-inclined than practice-inclined.
Transformer Language Model Mathematical Definition
Language models are big nowadays (pun intended). For those who are mathematically inclined, here is an interesting explainer written by Michal Ryszard Wojcik, a Ph.D. in Mathematics, under the guidance of one of the original co-authors of Attention is All You Need.
LovelyPlots: Matplotlib style sheets for Illustrator-editable figures
This is a cool repo. It provides a collection of style sheets to make matplotlib plots look nice while remaining editable in Adobe Illustrator. I see it as an excellent reference example of how to style up plots for your team, organization, and more.
Looking back at two years at Automattic and Tumblr
These are Vicki Boykis’ reflections on being a machine learning engineer at Automattic and Tumblr. As someone whose work necessarily straddles science and software development, many of her reflections resonate strongly with me. They’re likely valuable to those who want to jump into data science.