My 10 Tips for data-science undergrads: Part 1 - life hacks
Updated: Jun 26
[the opinions expressed here are entirely my own]
A letter to my 2021 undergrad students.
Since this is a day for celebration, and since I do believe in long term relationships between professors and students - I've prepared for you a list of 10 important tips that helped me during my career so far. The list, of course, is what I found to be helpful for my career path, and so I gladly share it with you. You definitely earned the right for a few shortcuts after a year and a half of learning through Zoom, while facing so many challenges due to COVID19.
Find a TA/teaching position in academia, and stick to it for many years after you graduate.
Story-telling is a vital skill for data practitioners. Nothing gets you motivated to tell a compelling story and prepare well more than standing in front of 100 students, while competing with their smartphones for attention.
Keeping in touch with academia will force you to get up to speed with advancements in the field.
Students are a great source of talent for your teams, once you become a manager.
Bright young minds, combined with a hunger for laddering up, keeps you sharp and gets you to constantly think of novel ideas.
Complete an MA thesis before the age of 30. (PhD if you wish - before the age of 35).
The worst things that happened to me during my thesis (and PhD dissertation) - prepared me the most for the hardest problems I tackled in my day to day work. During my MA I discovered what I don’t really like (theory), and discovered what I’m pretty good at (data).
Take your time, while you are young, to bang your head against the wall facing research questions no one tackled, or thinking you discovered a Nobel-Prize-winning finding, only to discover a day after on Google Scholar a similar finding, based on a paper from 1979 (true personal story).
Take the time to make research mistakes. Academic supervisors and Professors will be much more patient and tolerant than your future bosses (or the people who will interview you).
Don't linger to long to try it outside of academia (but do start in academia, while your mind is sharp and open).
Work for a tiny startup before the age of 30.
There is nothing wrong with working at a big corporation (I love it. My last two employers - Google and Outbrain, are the best places to work in the world imho).
However - the daily struggles of startups will forge you and will get you to admire everything that big corporations bring to the table.
It will get you to connect with many other startup employees in the business. Some of them will become your co-workers some day.
And it will forge your entrepreneurial spirit, thinking outside of the box, which big companies appreciate more and more and encourage people to bring as part of their skillset.
Work in big tech before the age of 40.
I like this quote - “There are two types of people in the world - those who divide people into two groups, and those who don’t”.
But seriously - there are mainly two types of people in the world - builders and farmers. Builders build from scratch end to end, usually patchy/imperfect solutions. They prove a point. Farmers, on the other hand - cultivate, nurture, set processes and think long term (ie - build for scale). You are much less likely to meet and learn from Farmers in small startups, while big companies have both.
Find mentors in your profession who are not only your boss.
My best mentors were (are) my bosses. I choose my positions based on that.
However - having also a mentor which is not your boss (not even from the same company) can help you get a fresh perspective, ideas and advice from the outside. Some of the best advice I received for my d2d jobs came from people in the industry that I admire and know me well enough to understand what is the right approach for me in challenging scenarios.
Make your colleagues your friends. Nothing much to say about that. Any industry is a small industry.
Make PMs your BEST friends, and understand how they think.
The difference between a tech company and a traditional company is becoming quite blurry (think AirBnB, Uber, SoFi and many others). Product managers are critical to how any org these days operates and develops.
Using cutting edge algorithms and ML can get you to lift the performance of a feature by 10’s of %s.
On the other hand - changing the product that is placed at the hands of an end-user - can lift performance by 100’s if not 1000’s of percentages. So we as Data Scientists should be humble when talking to PMs, as they control the resources, and they, in most cases, hold the future of a product in their hands.
Read quality content and pay for it. Take top-notch classes by top-notch professors/practitioners and pay for it (or better yet - ask your employer to pay for it. It is a sound investment).
Remember - if you are using a free product - then the product is you.
It is important to browse Social Media and scan new sites, but keep the right mix. For every minute you consume free content - make sure to consume quality content that costs money, for an additional 1m.
Time is money - use Audible for non-fiction (you can listen in X1.5 speed, and you don't get tired), Pay for Reading the Economist or other premium content (take it with a grain of salt though, but at least those folks pay someone to fact check their data), etc.
Keep notes while reading/listening using Google docs. You will be shocked how much you return to insights that you write in your notes.
Spend 30m on Coursera every day, learning something about ML, code and technology.
Data Scientists are researchers by profession. It means you constantly need to be up to speed with innovations in the field, or you will perish.
Atomic habits (learn 30m per day) will get you great results.
Adapt. The professions of Data Science and Product Analytics are evolving constantly.
Every Coursera graduate knows how to run k-means in NumPy, do Random Forest for “Iris” dataset, and do funnel visualizations with Florence Nightingale’s coxcomb chart.
There are more and more out-of-the-box algorithms and auto-hyper parameter tuning tools. Unless your work is around developing new algorithms for existing problems - it is more likely you will need to work on using existing algorithms for new problems.
Data practices will be integrated in any profession (Health, Education, Journalism, Sustainability, whatever). The algorithms will be quite the same, but it is about the last 20%, vertical-specific of the work which will make the difference.
Data folks will need to be more rounded, being able to carry e2e development work - from pitching an idea, conducting the research, modeling, bringing novel ideas for product development and landing the feature.
My advice - find your specific trait that makes you unique as a data scientist. All are not created equal and ds is becoming a rounded profession and interdisciplinary.
I practice what I preach - I am an Economist by profession, during my PhD I focused on researching incentives with data. That is what I am trained for. Even though I have much criticism for the field that I am coming from, I look at everything in the data field through the magnifying glass of an Economist (link). I was fortunate enough that Marketplace design is THE field these days.
Golden tip - acknowledge that most successful career moves are 90% luck and 10% skills. However - you can help luck by continuously investing in your skill sets, making learning a way of living, and making good friends with those you work with, those above you, below you, and around you.
Happy summer holiday dear students. Keep in touch.
IF you liked this post - share with someone who can benefit from it. Even if they are executives or your boss.