What can be added to a DS/ML engineer's portfolio/resume?

Question

Career in IT

What actually can be added to your resume, github, so that it attracts an employer or at least increases the HR.
MB are there any free (well, or not expensive) certificates that you can attach to fill yourself with value? What to upload to github (do they even pay attention to it, for example, if I have jupiter notebooks with content on the topic there)? What other ways are there (codeforces, kaggle competitions)?

Answer 1

A normal employer will not be attracted by certificates or laptops. A normal employer will be primarily attracted to the content of your previous projects. Thousands of practically copies of tasks walk around the network, often absolutely
useless from the point of view of real application, heroic rewritten from one laptop to another. And what's the point?
The employer will undoubtedly be attracted by your experience in solving practical problems in his subject area. I emphasize - "practical" and in "his" (or related) field. This means that you do not have to understand the nuances for months and the return from you will be fast enough. This factor is much more significant in DS than even in SW development or system administration. The trick is that it is in them - and in fact in the ability to translate the language of the subject area into the DS language - that a huge share of the "value" of the analyst lies. And a cool DS specialist in speech analysis may at first be practically helpless in solving problems in the field - for example, X-ray image recognition. Until it "drives" deep into the subject area. Although in fact the instruments are very close.
Certificates - all the more free, the thing that says only that you have spent a certain number of hours studying the subject - is commendable, but not impressive. Moreover, there are quite a lot of such certificates now.
Your success on Kaggle or similar contests can slightly increase your ranking - but you need to be careful here. Because there is a big difference between competitive solutions and real projects of real business. It's like "olympiad programming" and real software development experience.
By the way, your (confirmed) participation in some open projects looks interesting - at least it says that you know how to work in a team. Perhaps - success in some hackathrons, if you can beautifully describe what and how you solved there, even if you did not achieve victories there.
Well, something like this.
PS And yes, I'm talking about Data Analysis and Machine Learning.
"DS/ML engineer" is a completely different story, although many (and in the broad masses - most) do not understand this.

Answer 2

It is enough to arrange 2-3 of your projects in the form of a problem statement (text and images) and describe the approach to solving it (without formulas and code, only text and links to Wikipedia).
The result of achievements - to show in a video clip (the process itself) or in the final tabular / graphic data.
Bonus - you can attach a jupiter laptop (if you see fit to show all the logic and code).
The best: step-by-step visualization of the entire work in the form of a presentation (problem statement, algorithm execution, result), including step-by-step operation of the algorithm (step by step, as in debugging) with signatures and formulas where necessary.
That is, the "live book" or "live portfolio" mode .
And there must be a "live demo" link so that they can "play around" with their input data and make sure that everything works in reality!
(and not just in the pictures in the presentation!)
This is what will immediately show and prove your skills and the veracity of the words in the presentation.