Data Science #2.1: Interviewing - The First Screen

So on my data science journey, I have to jump ahead a few steps to the interview process and then we'll circle back to the skills you need to know to make it in data science.

Firstly, the interview process is chaotic and trying on everyone. Everyone. Even if you're kicking ass and taking names, you're still faced with essentially being on a stage all the time and facing decision pressure. If you're not getting 10 callbacks a day either, then you are facing crippling self-doubt (my favorite!). If you have friends applying for jobs, please don't ask them about it - it's rude. Let them tell you because, trust me, they will be dying to tell you once they have a signed offer letter.

For data science, the interview process is extra special - you're expected to demonstrate your brilliance at everything ranging from math puzzles to software engineering skills (i.e. build a linked list in Python). It's also expected that you're likeable ("noooo really?") and keep a blog where you're charming personality shines through (OK, but they definitely will check your blog).

So what happens when you finally get that magical callback or email where they (actually!) say "Hey, we read your resume and it was awesome! We'd like to talk to you more!"? You get a first screen - usually this is with a recruiter for bigger companies like Facebook, Amazon, etc. They just want to see if you're human and likeable. I have some good tips here:

(1) Before you get on the phone with that recruiter or data scientist, you better research that company fully. Maybe it will take you and hour, maybe six hours, but you need to know some facts about that company, understand their product, and get a sense for company direction. For example, if you're talking to Facebook, it's probably good to have a Facebook and understand how Facebook makes it's money and how you may be able to facilitate that process.

Put your thinking hat on and come up with an original
 idea (or at least an interesting one)
LoProTip: Think of a problem that company has and a way you could solve it. If you can, bring this up in your initial screen. It comes across as extremely impressive and like you deeply care about the company's problems. 
Example: LunchBoxes.com (fake company) shows mostly pictures of youth-targeted lunch boxes. You mention how adults also buy lunch boxes and how they might improve their profits if they also include a few adult lunch boxes on their homepage. That's it!

(2) You need to smile. Even if they can't see you, you should aim to be smiling and chipper. I have had screens where clearly the recruiter was having a rough day and was bored, and then *Lois* appeared in their life, and by the end of the conversation, I could tell they were happier and re-energized. You know what? I get responses faster and second interviews sometimes straight to onsites with this kind of attitude.

Be happy to be at this interview that decides your fate!

LoProTip: Pretend it's a good friend of yours and you are just jazzed to be talking about possibilities for their company with them. Keep the focus on the company - it's what you have in common :) 

Example: I love Lunchboxes.com - I actually own a penguin lunch box I bought from you all! Even better is that I noticed when I bought it that your recommender system tended to recommend me to other penguin lunch boxes, when in reality, I was more likely to buy a porcupine one. I think I could help you all update your recommender system to better accommodate these types of customer interactions based on my experiences working with machine learning and natural language processing on my thesis, etc. etc.


(3) Be honest. This should speak for itself, but a lot of people lie during their screens and this makes you look awful. If you don't want to move to Chicago, don't tell the recruiter you'd be happy to move to Chi-Town (seriously? WTF are you expecting to happen?). Tell the recruiter that you are interested in the position, but also are curious if they have positions open outside Chicago. That way, you're on the same page.

LoProTip: You have to live in the location you're working. Getting a job is hard, but you make it harder on yourself if you lie. Put a positive spin on things, but ensure honesty. 
Example: Hey lunchboxes.com, I see on glassdoor.com that your typical starting salary for a data scientist is $100k a year. My base level is $120k - would you be able to accommodate this in the position you are hiring for? I am willing to negotiate for equity or vacation days for a lower base pay.

(4) A little repetitive but... don't talk too much about yourself. They need to know how you and your skills will fit into the company, but don't go on a monologue about your thesis or a class project you've done. That's boring. Keep it concise and relevant.

Remember your audience - they are intelligent but may
lack your brilliance in your niche field
LoProTip: Always keep the language in an initial screen to 1st year Physics student when describing projects tangentially related to data science. Everyone feels better when they feel like they understand and/or are learning. 
Example: "You want to hear about my thesis, lunchboxes.com? I work with the most cold boring plasma in space that surrounds Earth and shields us from solar activity. My thesis work involved discovering that this plasma was a bit more exciting than we thought and through extensive data analysis, quantitative analysis, and time series work, we uncovered that plasma waves were interacting with low energy ions and causing unexpected heating in the plasmasphere. Cool, right?"

So that's enough to get you through your first behavioral screen. Let's talk about tech screens soon :)





You Might Also Like

4 Comments