Catch our monthly updates- featuring new video docs and talks, new jobs at DVC, and must-read contributions from the community about MLOps, data science with R, and ML in production.
Welcome to the November Heartbeat! Let's dive in with some news from the team.
Did you know we're hiring for two roles in our growing team? We're looking for:
A Senior Software Engineer for the core DVC team- someone with strong Python development skills who can build and ship essential DVC features.
A Developer Advocate to lead the community, support contributors and new users, and create new content like blogs and videos about DVC and CML.
Here are a few reasons to consider joining us:
If you're interested, we'd love to hear from you about either role (and we welcome referrals if you know a good candidate)!
We're continuing to develop our video docs, and now half of our "Getting Started" section has video accompaniments. Check out our latest release on data access with DVC:
This video covers functions like
dvc import, and the DVC Python
We took a quick break from releasing videos during the US election week, but look out for a new video on our YouTube channel about model testing with continuous integration! Subscribe to get alerts whenever we have something new :)
As usual, there are plenty of remote meetings on our schedules:
HealthData Bootcamp is a weeklong intensive for all things biomedical data science. Dmitry and myself (Elle), plus DVC Ambassadors Mikhail Rozhkov and Marcel Ribeiro-Dantas, will be presenting lectures and workshops about MLOps throughout the week!
I'll be leading a hands-on workshop at the Toronto Machine Learning Society Annual Meeting. It'll cover how to get started using Continuous Machine Learning(CML) with GitHub Actions- register here, and be sure to reserve your spot in the workshop.
This week, I have another talk at PyData Global about CML. PyData Global is online for the first time ever and promises to be a great gathering for Python-using data scientists in industry and academic research alike.
Here are some of our favorite happenings around the MLOps community this week.
Goku Mohandas, founder of Made with ML, announced plans to release a new online course about putting ML in production. The curriculum will cover everything from experiment tracking to deploying and monitoring models in production, and you can expect DVC to be included! Keep an eye on Goku and Made with ML on Twitter for updates.
🔥 Putting ML in Production! We're going to publicly develop @madewithml's first ML service. Here is the broad curriculum:— Goku Mohandas (@GokuMohandas) October 13, 2020
- 📦 Product
- 🔢 Data
- 🤖 Modeling
- 📝 Scripting
- 🛠 API
- 🚀 Production
More details (lessons, task, etc.) here: https://t.co/xmMm9XGK9j
Thread 👇 pic.twitter.com/T0uLPb2QbR
Dr. Larysa Visengeriyeva, creator of the top-notch "Awesome MLOps" GitHub repo, and DevOps expert Anja Kammer wrote a must-read essay about CI/CD for ML (note: it's published in German; I used Chrome's built-in translation to read in English).
The blog covers key concepts like continuous integration, deployment, and training with ML, as well as practical approaches and sample architectures.
Also, there's some cool art.
Another blog on our radar: Sean Lopp at RStudio made the first known blog about a CML report with a ggplot! Using RStudio's GitHub Actions for R and CML, Sean built a sample data science workflow that runs automatically in GitHub Actions on a push. He reports on some pros, cons, and areas for future development to make R language data science easy to automate.
Finally, developer Petr Stribny wrote about how to version big files in a Git project with DVC. It's a short-and-sweet guide to getting started, and if you're trying to decide if DVC is for you, this is worth a look.
To wrap it up, here's a kind tweet that we really like. It's always good to be mentioned in the same tweet as some of our heroes :)
Thanks for reading this month!