envd biweekly newsletter

[aseaday]

Could we open a newsletter for envd users or watcher.
I had some ideas for the content:

• For Students and ML/AI beginner: a kaggle problem or kaggle like ML/AI/Data Science Problem with its tutorial(Juypter Notebook).
• The updates and new user story of envd.
• Something new about ML/AI/Data Science and MLOps

[gaocegege]

Some suggestions:

• For investors and other MLOps companies: Introduce one MLOps company or its product in the bi-weekly newsletter. About its business model, its open-sourced projects, what problem they want to solve.

[gaocegege]

Here is an example https://www.linkedin.com/pulse/issue-204-ml-engineer-alejandro-saucedo/

[gaocegege]

Here is another example https://mlopsroundup.substack.com/

[gaocegege]

Maybe we can use https://www.getrevue.co/

[kemingy]

I wouldn't say I like this one because:

• its tracker or something else is not HTTPS, which will be blocked if the user browser has strict checks (like me)
• owned by twitter, but twitter may not be stable, you know what happened

[aseaday]

Overview of the newsletter

• Publish every two weeks
• Target audience:
  • envd users and fans
  • AI/ML Engineers(Not Researcher)
  • Students and ML/AI beginner
  • Companies's Tech Offical and insteresting investors
• Goals:
  • Most Important: Gather MLOps or AI/ML engineering fans together and build a neighborhood for them.
  • Promote our influence and exposure in ML/AI community especially in MLOps area.
  • Become the top 1 newsletter in MLOps area

What content we may publish

envd summary

A short essay describing the activities of

• envd community meetings notes
• envd development changelog
  But it should be orginized to be a in-depth article to clearly state changes and thinking of envd.

MLOps Radar

Introduction of MLOps trending:

• new player or tech of MLOps
• important change of MLOps
• new methodology/thinking of MLOps

The MLOps Radar is to give a whole a picture of what happended in last two weeks. It aimed at AI/ML Engineers, Companies's Tech Offical and insteresting investors. So it should't cover too much tech details but qualitative analysis matters.

Highlights for ML/AI

In this part, we will catch the best practices or engineering experiences in AI/ML area. such as:

• Synthetic Data Generation: How to Protect User Privacy
• Multi-cloud and cloud-agnostic computing practices

It is for AI/ML engineers to help them build better AI/ML applications. This part will have more details and dig into the core of tech.

envd camp

Intersting problems or challenges for students and beginners. We'd better choose some interesting contests such as kaggle or famous tools like xgboost. It worth considering because beginers have a strong demand and it also will bring a lot user search exposure. Here is some examples:

• Kaggle linear regression all-in-one solution
• 10mins to set up your stock prediction environment

We also can actively collobrate with other ML/AI/Data projects to make it sexy.

How to organize the edition work

In order to create content regularly and at a high quality. We'd better choose a chef editor. Maintainers or contributors in envd community take turns to be the edior.

The editors can call submissions from all envd community memebers.

Reward for Contributors

I think we could have some kind of reward for contributors to submit articles qualified. A gift card and a meal?

How we publish the newsletter

We can hold four ways to publish the newsletter

• substack: for common users who like beautiful UI and substack have the discussion function.
• medium/blog: search engine optimization
• Wechat: for chinese users

[gaocegege]

LGTM, let's give it a try!

[AlexXi19]

I personally would love to read MLOps or ML related articles from an industry standpoint.

Genuine question, are there really any ML/AI fans or students who play around with models in their free time? If so, what are some of the things they play around with? Because I imagine it would be hard to play around with large deep learning models.

[aseaday]

Thanks for you reply. I mentioned students because I spend a lot of time help student (non-cs) set up their environment when I was TA. We could prepare some basic projects for students.
As for large deep learning models, the only problem is the cuda equipment with the help of envd. I think we can solve it by providing a envd cloud service LOL in the future like colab or we could embed envd in colab service.

[AlexXi19]

Ah yes, that is indeed a use case I have thought about.

envd cloud service
😏

[gaocegege]

We need a name for the newsletter.

It's your turn to show up your ability.

[gaocegege]

Here are some ideas to get you started:

• MLOps roundup from
  @mlopsroundup

• The ML Engineer

[kemingy]

How about "42ml"?

• 42 is a common random seed. Mainly because it's the "Answer to the Ultimate Question of Life, the Universe, and Everything"
• ml is short for machine learning

[xieydd]

How about "MLOops", o͝ops hh.