Interesting stuff I found in February
Dashboarding with Python, transparent compute infrastructure, and more deep learning models. Plus co-op positions with the DSAI team at Moderna!
Hello, fellow datanistas!
It's March! How time flies. I hope you all are doing brilliant and excellent. Here's another edition of the DSPN newsletter - this time with exciting stuff across the data science space.
Dashboarding in the Python Ecosystem
In 2019, I wrote an essay on the Python dashboarding landscape. 4 years on, my friend Sophia Yang, Sr. Data Scientist at Anaconda, brings us an updated look at the landscape of tooling. How things have changed! While I no longer build user-facing products in my role at Moderna, building dashboards can be part of the job expectations at other places. Knowing which dashboarding tool fits well into your organization's technical stack can help a ton. Check out her essay for tons of valuable advice!
Computation without worrying about infrastructure
Built by Erik Bernhardsson, Modal promises to run things in the cloud - burst compute, serving APIs, and more - without considering infrastructure. I read through the docs, and it looks promising! Check it out here.
Deep Learning Playbook
A bunch of Googlers released their deep learning playbook. Lots of practical advice in there about how to train and tune deep learning models from people who have accumulated more than decades of experience doing so. Check it out here!
Google releases MusicLM, a model for text-to-music synthesis
This feels massive! Being able to describe the kind of music you'd like and get back high-quality audio is impressive. Also raises lots of legal questions. The model has yet to be released for public use because it has one big flaw: incorporating copyrighted music about 1 percent of the time. Check out the project's GitHub page here. TechCrunch's early coverage also hits most of the main talking points while being more accessible.
From my collection
Moderna Data Science & AI co-ops
Moderna digital is hiring! We have co-op positions for July-December 2023. Check them out here.
I built an app that uses GPT-3
I tried out the OpenAI GPT-3 API and used it to build a translation tool. My biggest lesson from this exercise? GPT-3 can be a pretty awesome data engineer. Curious to see more? Check out my latest blog post for more detail!
And that's a wrap!
I hope you found this newsletter edition helpful, informative, and valuable! If you did, I'd love to hear back from you - words of encouragement keep me motivated to continue sharing! (You can drop me a note by commenting on Substack or by finding me on Mastodon/Twitter/LinkedIn.)