In the past few years, I have met up with a lot of employers and conducted interviews for training program. Through the conversations and interviews and seeing the end results, I thought I will share more on how to prepare for your resume and even the interviews for a data science role. Most of the tips are for people who want to enter into the data science profession with a "green" background. I cannot promise results but I hope it can help those who are passionate about data science. The tips given are really for those who are passionate in it as it requires a lot of effort.
The title explains itself but let me take the opportunity to explain it further. Often I was asked "Can I use my school project as part of the project portfolio?" Now here are some considerations I have if one were to showcase their school projects.
Firstly, most of these school projects are guided (hopefully) and work in teams. Its very challenging to differentiate which part of the project is done by who. What I can only infer is that, based on the results of the project, whether the team is functional or dysfunctional.
The best project to showcase is done outside of the curriculum, during one's free time because it shows that the person is passionate in data science and willing to spend their free time on it. I can also attribute whatever that is done in the project to the interviewee. But that is after I have asked a few more questions on the project to ascertain it.
Do have keywords (such as the machine learning models used, the model training process etc) used to explain the project but be prepared to explain those keywords especially when it is a technical interview. I tend to ask interviewees to explain those keywords and the explanation has to be at a level that the layman can understand. Well, if one cannot explain it to the layman, it means one still does not understand it completely right?
If one is to work in the data science or even AI, being able to work in a team matters tremendously, regardless of being in a leadership or team player role. So it is important to showcase any team projects and also the impact achieved, preferably quantify the impact so the interviewer can build a good mental impression. The impact will give me some information on whether the interviewee can work in a team or not. I also tend to ask the interviewee to share more about their experience in at least one project, so as to ascertain if he/she can work in a team.
Maths & Statistics Background
I like a good grasp on the level of mathematics and statistics that the interviewee has. It can be inferred from the module grades, projects and tools that they used.
Module grades does help to ascertain the level. I usually look at the whole portfolio of mathematical modules that were taken as a whole to ascertain the level of maths and stats background. I do give chance for those that have mediocre grades but I will definitely ask the interviewee why the mediocrity. I do ask the interviewee about what they like and do not like about maths and stats to determine if he/she can work with the mathematics required in data science and AI.
The projects and tools does help to infer the maths background. This is seen through the machine learning models they have used, how they implement it and why they implement it in a particular manner. I may ask what were the challenges they faced during these projects, the reason a particular solution was chosen and as much as possible relate to the mathematics behind.
Where possible, do showcase any codes written, especially if the codes are written for data science projects. Otherwise, other languages are welcomed, not necessary must be those common languages used in data science (R, Python, Scala).
Make sure it is well-documented. Well-documented meaning there is a good description on what the code is doing, why the codes need to be written in such a way, why the code is implemented etc. The main objective is for the interviewer to understand the thought process interviewee have gone through in writing codes and deriving insights from the project. Documentation is very important in order for data science to be reproducible, interpretive and accountable. Showcasing the thought process is a very important consideration for the interviewer to determine how much autonomy can be given to the new hire to get meaningful results from the project.
Thus a well-documented code is very important, as it can indicate the level of knowledge, skills and thinking of the potential hire.
The above points are what I have gathered interviewing for training program and talking to numerous employers. I hope the points shared will help you to construct a more "attractive" portfolio and resume to your potential employers and also be well-prepared for possible interview questions to come.