There was a “fireside chat” at rstudio::conf 2018 about R in industry, where Eduardo Ariño de la Rubia said something that resonated well with me. The panel was discussing about how to hire data scientists, and Eduardo shared some of his criteria of choosing candidates. I’m not sure if I paraphrased them correctly below.
You need to have a Github presence (I’m sure it does not have to be Github, e.g., Gitlab should also be fine). It is hard to imagine that a data scientist does not use version control.
You must have given talks at local meetups or conferences. Communication is a key part of data science. If someone invites you to give a talk, it is also evidence of the importance of your work.
You need to have a few (data-related) projects of your own, without anybody asking you to do them. These projects show that you are intrinsically interested in data science.
The above evidence speaks louder than your CV. Everyone can say that he/she is good at R, Python, machine learning, and so on, you name it, in the CV, but you need something more concrete to show your capabilities and curiosities.
I have never been involved with hiring people, but if I do, I’d check if they have personal websites or have written anything nontrivial.
- Update on 2018-02-16
After I read the comments by Rebecca Robare and sleight82 below, I realized that I made a mistake. I think Eduardo’s words resonated well with me partly because I would be a good candidate meeting the above criteria (I use GIT every day, I have given numerous talks, and I’m highly self-motivated). In other words, I unconsciously chose to write down things that were good for myself. In conclusion, I think this post was biased, and I’d welcome more managers to share their criteria.