5 Essential Lessons for Junior Data Scientists I learned at Spotify (Part 2) | by Khouloud El Alami | Jun, 2023

And maaaan the time it would have saved me had I learned this sooner rather than later.

Picture this:

I’m working on a causal inference project (field of statistics that aims to identify the causal relationship between variables based on observational data). I had taken the course in college, but I couldn’t seem to recall much from it (maybe I was sleeping again this time). So anyway I’m working on this project, tackling new exciting concepts to tickle my brain just like I like it.

I’m asking my closest peers for advice, diving back into past projects, hoping to find some inspiration, learnings, tips, god’s word… really anything that can help me. So yes, this one is also an important skill to have under your sleeve:

Tracing back into past resources should always be a first step when starting new projects

but I’ll dive more into this in another story.

So I’m doing my research, investigating internal and external resources and I do find myself butting heads with causal inference (normal, that’s a tricky one). I do all the right things (or so I thought). I go on and on working on this project, and after some time… I come up with the great idea of asking other data scientists for clarification on a concept.

By doing that, I obviously provide more details on my project. When… out of nowhere… the word of a heaven-descended causal inference expert shone down upon me… to nicely let me know that I was… barking up the wrong tree. My whole methodology was off track because I was comparing two populations of users that cannot be compared, which throws off the whole analysis.

And there, my friends, that’s how weeks time of work earns a one-way ticket to the trash! (Well not completely, because this becomes a core memory in your brain and a nice lesson to live by)

Which brings us to lesson #4 — Learning how to overcome challenges on your own is important in developing your critical thinking and problem-solving skills. But learning to ask for help whenever needed is also as important. People go right to ChatGPT now for help instead of giving it a shot on their own, and this ultimately prevents them from learning the right skills.

Asking for help from the right people has valuable benefits

1. Getting firsthand guidance from experts in whatever you’re working on

Clearly this can only but give you a boost: a) for your project and b) for your skillset. Remember, you might be a puppy but you’re also playing in the same sandbox as experts in the field, so don’t forget to ask for guidance when needed (unless you’re planning on staying a puppy for a bit longer, then that’s another story).

How can you do that?

1. Look for past projects where what you are working on has been implemented or researched. Then reach out to the people who worked on them. Chances are, they will definitely provide you with valuable information and help you identify potential inconsistencies.
2. Send a message on the Slack/Teams channels dedicated to the problem/technique/feature/product area etc… you are working on, eg. #causal-inference or #data-science (with a wider net, someone will definitely take the bait)!

Everyone is always happy to help. After all, they’ve been there themselves.

2. Spare you the frustration of realising you’ve been doing things wrong

This will ultimately also save you from the stress of having to work extra to make up for your past mistakes… because now, you’re also running late on your schedule.

Just don’t forget to at least give it a try first. If you find yourself getting stuck longer than you should, then you know it’s time to seek for a helping hand.

If you’ve gotten all the way through here then you definitely deserve one extra cookie for sticking around. Thank you for reading me.

Brace yourself, now I will bestow on you my ultimate cookie for this ride.

No one expects you to be an expert from day 1 or even day 100

Even in Tech! Working alongside experienced people is a unique opportunity to grow between good hands. However, it did take me some time to be fully okay with being the most unexperienced one in the room.

I looked up to the people I worked with, but I also subconsciously compared myself to them:

• It would take me much longer to deliver my work
• I would not always ask the right questions to explore, so it felt like my exploration was limited in scope compared to my peers
• I would struggle with juggling between more than 1 project at a time, while others felt like they were smooooth sailing through 5 tasks at a time

Yes, it might seem obvious, but it wasn’t that obvious to me at least. So if like me, at times you can be hard on yourself. Striving for the best is important. However, know it’s okay not to be the best when you’re just getting started.

I mean come on, you’re a puppy, no one expects you to be on the same level as fully grown wolves anyway. But you’re in the pack, and the pack does not leave its own behind. So don’t worry, you’ll howl too one day, it’s only a matter of time, and commitment.

Getting everything right from the beginning would actually not be the norm. Besides, surfers do not ride still waters, where’s the fun in that? Even Harry Potter didn’t get his Wingardium LeviOsa right from the get go.

So what can you do?

1. Avoid overdoing it by trying to imitate your seniors. Chances are, it will not feel natural, and people will sense it. Instead, ask questions without forcing it and be yourself
2. Reach out to other fellow juniors. Exchanging with other people who I could relate to definitely helped me get more perspective when I was struggling with this myself. It does not have to be other data scientists, any junior you’re closing enough with will do. Knowing that you’re not alone and finding support changes everything.

So last lesson: Cut yourself some slack and stop putting that extra pressure on yourself if you’re doing that. If not, then this lesson might not be for you, but it’s still worth to keep in mind and remember to be kind to yourself.